Methods, equipment, media and products for identifying intermittent faults in fiber optic gyroscope fusion splices

Abstract

This invention discloses a method, device, medium, and product for identifying intermittent faults at fiber optic gyroscope fusion splices. The method includes: synchronously collecting multi-dimensional parameters using multiple sensors deployed at key fusion splices of the fiber optic gyroscope; inputting the multi-dimensional parameters into a pre-trained multimodal diagnostic model; after generating identification results for each dimension through the multimodal diagnostic model, fusing the identification results for each single dimension to obtain and output the final fused fault type and fused fault confidence level; when the fused fault confidence level is greater than or equal to a confidence threshold, triggering a multiphysics stress loading test on the key fusion splice and detecting whether the fused fault type can be reproduced during the test; if so, outputting the fused fault type as the intermittent fault diagnosis result for the key fusion splice. The technical solution of this invention effectively solves the problem of finding and reproducing intermittent faults, improving the accuracy and reliability of fault diagnosis.

Description

Technical Field

[0001] This invention relates to the field of fiber optic gyroscope fault detection technology, and in particular to a method, device, medium, and product for identifying intermittent faults at fiber optic gyroscope fusion splices. Background Technology

[0002] With the increasing application of fiber optic gyroscopes in aerospace, defense, and high-precision navigation, long-term operational reliability has become a core indicator. Faced with complex and variable environmental stresses such as vibration and temperature shocks, intermittent failures at the fusion splice points caused by micro-defects in the manufacturing process or material aging have become the primary hidden danger affecting the reliability of fiber optic gyroscopes. These failures are intermittent and difficult to reproduce, posing a significant challenge to fault location and root cause analysis.

[0003] In existing technologies, the detection of fiber optic fusion splices mainly relies on splice loss assessment upon completion and static parameter testing using an optical time-domain reflectometer. However, these methods can only obtain static performance indicators of the splice and cannot effectively capture its transient response and performance fluctuations under dynamic stress environments. For intermittent faults, existing technologies lack effective online monitoring methods and triggering mechanisms, resulting in insufficient evidence for fault diagnosis, making it difficult to accurately locate the fault and provide effective process feedback, thus hindering the quality improvement and technological advancement of high-reliability fiber optic gyroscope products. Summary of the Invention

[0004] This invention provides a method, device, medium, and product for identifying intermittent faults in fiber optic gyroscope fusion splices, which can achieve accurate diagnosis of intermittent faults in fiber optic gyroscope fusion splices.

[0005] According to one aspect of the present invention, an intermittent fault identification method for fiber optic gyroscope fusion splices is provided, the method comprising:

[0006] Multiple types of sensors are deployed at the key fusion splice points of the fiber optic gyroscope to synchronously collect multi-dimensional parameters of the key fusion splice points; among these, the multi-dimensional parameters include optical characteristic parameters, mechanical characteristic parameters, and thermal characteristic parameters.

[0007] The multi-dimensional parameters of the key weld points are input into a pre-trained multimodal diagnostic model;

[0008] After generating single-dimensional identification results corresponding to optical, mechanical, and thermal characteristic parameters respectively through the multimodal diagnostic model, the single-dimensional identification results are fused to obtain and output the final fused fault type and fused fault confidence.

[0009] When the confidence level of the fusion failure is greater than or equal to the confidence threshold, a multiphysics stress loading test is triggered on the key fusion joint, and it is checked whether the fusion failure type can be reproduced during the test.

[0010] If so, the fusion fault type is output as the intermittent fault diagnosis result for the critical fusion point.

[0011] According to another aspect of the present invention, an intermittent fault identification device for fiber optic gyroscope fusion splices is provided, the device comprising:

[0012] The synchronous data acquisition module is used to synchronously acquire multi-dimensional parameters of the key fusion splice point through multiple types of sensors deployed at the key fusion splice point of the fiber optic gyroscope; among which, the multi-dimensional parameters include optical characteristic parameters, mechanical characteristic parameters and thermal characteristic parameters;

[0013] The data input module is used to input the multi-dimensional parameters of the key weld points into a pre-trained multimodal diagnostic model;

[0014] The multimodal fusion decision module is used to fuse the single-dimensional identification results corresponding to optical feature parameters, mechanical feature parameters and thermal feature parameters respectively after generating them through the multimodal diagnostic model, and to obtain and output the final fused fault type and fused fault confidence.

[0015] The detection and reproduction module is used to trigger a multiphysics stress loading test on the key weld joint when the confidence level of the fusion failure is greater than or equal to the confidence level threshold, and to detect whether the fusion failure type can be reproduced during the test.

[0016] The fusion output module is used to output the fusion fault type as the intermittent fault diagnosis result of the key weld point when a fault of the fusion fault type is detected that can be reproduced during the test.

[0017] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0018] At least one processor; and

[0019] A memory communicatively connected to the at least one processor; wherein,

[0020] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform an intermittent fault identification method for fiber optic gyroscope fusion splices according to any embodiment of the present invention.

[0021] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement an intermittent fault identification method for fiber optic gyroscope fusion splices as described in any embodiment of the present invention.

[0022] According to another aspect of the present invention, a computer program product is also provided, including computer instructions that, when executed by a processor, implement the steps of the method as described in any embodiment of the present invention.

[0023] The technical solution of this invention involves simultaneously acquiring multi-dimensional parameters, including optical, mechanical, and thermal characteristic parameters, from multiple sensors deployed at key fusion splices of a fiber optic gyroscope. These multi-dimensional parameters are then input into a pre-trained multimodal diagnostic model. The multimodal diagnostic model first generates single-dimensional identification results corresponding to the optical, mechanical, and thermal characteristic parameters, respectively. Subsequently, the single-dimensional identification results are fused to obtain and output the final fused fault type and fused fault confidence level. When the fused fault confidence level is greater than or equal to a confidence threshold, a multi-physics stress loading test is triggered on the key fusion splice, and it is checked whether the fused fault type can be reproduced during the test. If reproduction is successful, the fused fault type is output as the intermittent fault diagnosis result for the key fusion splice. This novel intermittent fault identification method for fiber optic gyroscope fusion splices effectively solves the technical bottleneck of traditional detection methods in capturing transient anomalies and reproducing sporadic faults. Through the mechanism of multi-source information fusion and active stress loading verification, the accuracy and reliability of fault diagnosis are significantly improved.

[0024] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a flowchart of an intermittent fault identification method for fiber optic gyroscope fusion splices according to Embodiment 1 of the present invention;

[0027] Figure 2 This is a flowchart of another intermittent fault identification method for fiber optic gyroscope fusion splices provided in Embodiment 2 of the present invention;

[0028] Figure 3 This is a flowchart of another intermittent fault identification method for fiber optic gyroscope fusion splices provided in Embodiment 3 of the present invention;

[0029] Figure 4 This is a schematic diagram of the structure of an intermittent fault identification device for a fiber optic gyroscope splice according to Embodiment 4 of the present invention;

[0030] Figure 5 This is a schematic diagram of the structure of an electronic device that implements an intermittent fault identification method for fiber optic gyroscope fusion splices according to an embodiment of the present invention.

[0031] In this array, 10 is an electronic device, 11 is a processor, 12 is a read-only memory (ROM), 13 is a random access memory (RAM), 14 is a bus, 15 is an input / output (I / O) interface, 16 is an input unit, 17 is an output unit, 18 is a storage unit, and 19 is a communication unit. Detailed Implementation

[0032] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0033] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0034] Example 1

[0035] Figure 1This is a flowchart of an intermittent fault identification method for fiber optic gyroscope fusion splices provided in Embodiment 1 of the present invention. This embodiment can be applied to the diagnosis of intermittent faults in fiber optic gyroscopes that are difficult to reproduce. The method can be executed by an intermittent fault identification device for fiber optic gyroscope fusion splices. This device can be implemented in hardware and / or software and is generally configured in electronic devices.

[0036] Intermittent faults at fiber optic gyroscope splices can be understood as temporary connection failures or performance fluctuations at the splice point during fiber optic splicing or long-term use, caused by defects in the splicing process, environmental stress, or material degradation. These faults are triggered under specific conditions (such as vibration, temperature changes, or contaminant interference), manifesting as intermittent optical signal transmission, sudden increases in loss, or abnormal reflectivity. However, they may temporarily recover after the stress dissipates, recur repeatedly, and are difficult to reproduce stably through conventional testing.

[0037] Correspondingly, such as Figure 1 As shown, the method includes:

[0038] S110. By deploying multiple types of sensors at the key fusion splice points of the fiber optic gyroscope, multi-dimensional parameters of the key fusion splice points are collected synchronously; among which, the multi-dimensional parameters include optical characteristic parameters, mechanical characteristic parameters and thermal characteristic parameters.

[0039] In this embodiment, the key fusion splice point can be understood as the fusion interface connecting the fiber optic loop and core optical components such as the Y-waveguide in the fiber optic gyroscope. The optical transmission characteristics, mechanical stability, and thermal stability at this location directly determine the accuracy and reliability of the entire gyroscope system.

[0040] In this embodiment, since intermittent faults at the critical fusion splice of the fiber optic gyroscope are non-persistent and only briefly manifest under specific external conditions, it is necessary to implement synchronous real-time monitoring of multi-dimensional parameters around the fusion splice to ensure that corresponding parameter changes can be captured in a timely manner when intermittent faults occur. Specifically, sensors for monitoring optical, mechanical, and thermal characteristic parameters are deployed at the critical fusion splice, and continuous synchronous acquisition of multiple parameters is carried out using these sensor groups.

[0041] S120. Input the multi-dimensional parameters of the key weld points into the pre-trained multimodal diagnostic model.

[0042] The multimodal diagnostic model can be understood as an intelligent analysis model capable of processing multiple types of data simultaneously. In this solution, the model is specifically designed to fuse parameters from three different dimensions—optical, mechanical, and thermal—from the fiber optic splice, enabling it to discover hidden correlations between different physical quantities and effectively identify complex intermittent faults that cannot be determined by a single sensor.

[0043] In this embodiment, the obtained multi-dimensional parameters are preprocessed and organized into three types of data streams: optical characteristic parameters, mechanical characteristic parameters, and thermal characteristic parameters, forming a data format that meets the input interface requirements of the multimodal diagnostic model, ensuring strict synchronization and dimensional uniformity of different physical quantities in the time dimension.

[0044] S130. After generating single-dimensional identification results corresponding to optical, mechanical and thermal characteristic parameters respectively through the multimodal diagnostic model, the single-dimensional identification results are fused to obtain and output the final fused fault type and fused fault confidence.

[0045] Among them, the fusion fault type can be understood as the fault classification result comprehensively determined by the multimodal diagnostic model after collaborative analysis of multi-dimensional parameter features such as optics, mechanics, and thermal properties. The fusion fault confidence can be understood as a quantitative evaluation index of the reliability of the current diagnostic result by the multimodal diagnostic model. Its value depends on the consistency of evidence and the significance of features among the multi-dimensional parameters.

[0046] In this embodiment, the multimodal diagnostic model first processes the corresponding single-dimensional parameters through optical subnetwork, mechanical subnetwork and thermal subnetwork respectively. After each subnetwork outputs the preliminary diagnostic results, a dynamic weighted fusion strategy based on attention mechanism is adopted to adaptively adjust the contribution weight according to the data quality of each modality. Finally, a comprehensive diagnostic conclusion is synthesized through evidence theory.

[0047] S140. When the confidence level of the fusion failure is greater than or equal to the confidence threshold, trigger a multiphysics stress loading test on the key fusion joint and check whether the fusion failure type can be reproduced during the test.

[0048] Among them, multiphysics stress loading test can be understood as an active fault reproduction technology. Its core idea is to actively induce potential intermittent faults by simulating the complex working conditions that fiber optic gyroscopes may encounter in practical applications.

[0049] In this embodiment, based on the confidence assessment of the diagnostic results, when the fusion confidence exceeds a preset threshold, it indicates that the diagnostic results of a single mode have formed mutual corroboration. At this time, the multi-physics coupling loading device is activated through the integrated control platform. While maintaining the continuity of optical monitoring, broadband vibration excitation and rapid temperature change stress are applied simultaneously to construct service environment conditions that match the fault characteristics.

[0050] S150. If a fault of the fusion fault type can be reproduced during the test, the fusion fault type is output as the intermittent fault diagnosis result of the key fusion point.

[0051] In this embodiment, the optical response signal collected during the stress loading process is compared with the target fault feature template in real time to calculate the dual indicators of waveform similarity and duration matching. When the correlation coefficient exceeds the judgment threshold and the fault duration falls within the typical range, the fault is judged to be successfully reproduced. At the same time, the critical stress parameters that trigger the fault are recorded to form a complete diagnostic report containing the fault type, reproduction conditions and feature parameters.

[0052] The technical solution of this invention involves simultaneously acquiring multi-dimensional parameters, including optical, mechanical, and thermal characteristic parameters, from multiple sensors deployed at key fusion splices of a fiber optic gyroscope. These multi-dimensional parameters are then input into a pre-trained multimodal diagnostic model. The multimodal diagnostic model first generates single-dimensional identification results corresponding to the optical, mechanical, and thermal characteristic parameters, respectively. Subsequently, the single-dimensional identification results are fused to obtain and output the final fused fault type and fused fault confidence level. When the fused fault confidence level is greater than or equal to a confidence threshold, a multi-physics stress loading test is triggered on the key fusion splice, and it is checked whether the fused fault type can be reproduced during the test. If reproduction is successful, the fused fault type is output as the intermittent fault diagnosis result for the key fusion splice. This novel intermittent fault identification method for fiber optic gyroscope fusion splices effectively solves the technical bottleneck of traditional detection methods in capturing transient anomalies and reproducing sporadic faults. Through the mechanism of multi-source information fusion and active stress loading verification, the accuracy and reliability of fault diagnosis are significantly improved.

[0053] Example 2

[0054] Figure 2 This is a flowchart of an intermittent fault identification method for fiber optic gyroscope fusion splices according to Embodiment 2 of the present invention. This embodiment is an optimization based on the above embodiments. In this embodiment, the operations of "optical characteristic parameters, mechanical characteristic parameters, thermal characteristic parameters" and "synchronously collecting multi-dimensional parameters of the key fusion splice by deploying multiple types of sensors at the key fusion splice of the fiber optic gyroscope" are refined. Among them, optical characteristic parameters include light intensity change and back reflectivity; mechanical characteristic parameters include axial strain value; and thermal characteristic parameters include temperature distribution data and temperature gradient data.

[0055] Light intensity variation can be understood as the degree of "brightness" attenuation of the light signal after passing through the fusion splice, directly reflecting whether there is blockage or leakage at the splice. Back reflectivity can be understood as the intensity of the light signal reflected back after "colliding" with an obstacle at the fusion splice; an abnormal increase in back reflectivity is like an echo suddenly becoming clear and piercing, indicating that there may be cracks or contamination at the fusion interface. Axial strain can be understood as the degree of fiber deformation at the fusion splice; abnormal changes in axial strain are like detecting minute deformations in bridge cables, sensitively reflecting whether external mechanical stress has damaged the fusion splice. Temperature distribution data can be understood as the temperature sensing along the axial direction of the fusion splice, locating localized poor heat dissipation caused by process defects. Temperature gradient data can be understood as the rate of temperature change per unit distance at the fusion splice, revealing the risk of thermal stress concentration due to material mismatch.

[0056] Correspondingly, such as Figure 2 As shown, the method includes:

[0057] S210. The light intensity change at the critical fusion splice is collected by fiber optic couplers and photodetectors set up upstream and downstream of the critical fusion splice.

[0058] In this embodiment, non-destructive monitoring of optical signals is achieved by employing optical splitting coupling technology. Specifically, a specific ratio of fiber optic couplers is connected upstream and downstream of the fusion splice, respectively, to divert a small portion of the optical power in the main optical path to a high-response-speed photodetector. By comparing the differences in optical intensity values ​​between the upstream and downstream locations in real time, the instantaneous changes in transmission loss at the fusion splice are accurately captured.

[0059] In a specific example, 99:1 fiber couplers are connected 5cm upstream and 5cm downstream of the fusion splice, shunting only 1% of the optical signal to the detector to ensure that the gyroscope's normal operation is not affected. An InGaAs (Indium Gallium Arsenide) photodetector is used to monitor the intensity of the incident light (upstream) and transmitted light (downstream) in real time, with a resolution of 0.001dB (decibel) and millions of samples per second. When a sudden change in the upstream and downstream light intensity difference is detected (for example, a certain type of gyroscope can be set to a sudden increase of 0.5dB lasting for 2ms), it can be identified as an abnormal loss event at the fusion splice.

[0060] S220: The back reflectivity of the critical fusion splice is collected by an optical fiber circulator and an optical power meter set up upstream and downstream of the critical fusion splice.

[0061] In this embodiment, the directional transmission characteristics of the fiber optic circulator are used to collect reflected signals. By deploying a three-port circulator upstream of the fusion splice, the incident light is directed through the fusion splice. At the same time, the weak reflected signal generated at the fusion splice interface is separated to a dedicated detection channel and quantitatively analyzed by a high-sensitivity optical power meter, thereby realizing dynamic evaluation of the fusion splice quality.

[0062] In a specific example, a three-port fiber optic circulator is deployed upstream of the fusion splice. Utilizing the minute reflections at the splice interface (normal reflectivity <-70dB), the reflected light is directed to a high-sensitivity optical power meter. If the reflectivity suddenly increases by more than 5dB and lasts for more than 10ms, or if the reflectivity suddenly increases by more than 5dB more than three times within one minute, an alarm is immediately triggered.

[0063] S230: The axial strain value of the critical fusion splice is collected by fiber optic grating sensors arranged at preset intervals on both sides of the critical fusion splice.

[0064] In this embodiment, mechanical deformation monitoring is performed using the wavelength encoding principle. A series of fiber optic grating sensors are arranged at precise intervals along the fiber axis. When mechanical deformation occurs at the splice, it will cause a change in the grating period, which in turn changes the reflected wavelength. The wavelength drift can be monitored by a demodulation device and converted into the corresponding axial strain value.

[0065] In a specific example, fiber Bragg grating sensors (wavelength 1530-1565nm) are arranged axially at 10mm (millimeter) intervals at both ends of the fusion splice to measure axial strain in real time (accuracy 1). When the strain difference between adjacent sensors is >20 At the same time, the location of the fault is determined by combining light intensity data, among which, The unit of measurement for the dependent variable.

[0066] S240: Collect temperature distribution data and temperature gradient data of the critical fusion splice by using distributed temperature-measuring optical fibers wound around the surface of the protective sleeve of the critical fusion splice.

[0067] In this embodiment, temperature field reconstruction is achieved based on optical time-domain reflectometry. A specially coated temperature-measuring optical fiber is attached to the surface of the fusion splice protective sleeve in a specific winding manner. By injecting optical pulses and analyzing the intensity change of backscattered light, the temperature distribution curve along the length of the optical fiber is obtained by inversion, and then the temperature gradient data is obtained by differential operation.

[0068] In a specific example, a distributed temperature-sensing optical fiber (based on the principle of Raman scattering, with a spatial resolution of 0.1 m) is spirally wound around the surface of the fusion splice protective sleeve, and the temperature distribution is scanned every second. If a local temperature rise rate > 0.5℃ / s (degrees Celsius per second) or a temperature gradient > 0.2℃ / cm (degrees Celsius per centimeter) is detected, the temperature distribution is monitored.

[0069] S250. Input the multi-dimensional parameters of the key weld points into the pre-trained multimodal diagnostic model.

[0070] Furthermore, based on the above embodiments, before inputting the multi-dimensional parameters of the key weld points into the pre-trained multimodal diagnostic model, it may also include...

[0071] Real-time acquisition of light intensity changes and back reflectivity at multiple reference fusion splices on multiple reference fiber optic gyroscopes;

[0072] If the light intensity decrease rate of the target reference weld point exceeds the preset decrease rate threshold within a preset time period, or the reflectivity increase rate exceeds the preset increase rate threshold within a preset time period, then continue to collect axial strain values ​​and temperature distribution data on the target reference weld point.

[0073] If the mechanical strain rate of the target reference weld point exceeds the preset strain rate threshold within a preset time period, or the temperature change rate exceeds the preset change rate threshold within a preset time period, then the full waveform of the reference multi-dimensional parameters of the target reference weld point is acquired at the preset acquisition frequency; otherwise, the reference multi-dimensional parameters of the target reference weld point are cached for a short time.

[0074] Timing analysis is performed on the reference multi-dimensional parameters acquired in full waveform acquisition or short-time buffer to obtain the reference fault type corresponding to each reference multi-dimensional parameter.

[0075] Multiple training samples are constructed using each reference multidimensional parameter and the corresponding reference fault type. The training samples are then used to train a pre-defined machine learning model to obtain a multimodal diagnostic model.

[0076] Generally, this embodiment first constructs a fault sample database covering various operating conditions. By monitoring a large number of critical fusion splices of in-service fiber optic gyroscopes over a long period, it continuously records basic optical parameters such as light intensity changes and back reflectivity. When a fusion splice exhibits fault symptoms that meet preset thresholds, such as abnormal light intensity attenuation or a sharp increase in reflectivity, within a specific time window, the point is automatically marked as a high-risk target, and extended acquisition of its axial strain value and temperature distribution data is immediately initiated.

[0077] Generally, after obtaining extended data, it is further determined whether the mechanical strain rate or temperature change rate exceeds the secondary threshold. If an abnormal strain rate (such as periodic deformation caused by vibration) or a sudden temperature change (such as local overheating) is detected, a high-speed acquisition mode is activated to capture full waveform data of the weld joint; if the triggering condition is not met, data is buffered at a conventional frequency. This multi-level triggering mechanism ensures the complete acquisition of fault characteristics while effectively reducing the data storage load.

[0078] Generally, after obtaining extended data, it is further determined whether the mechanical strain rate or temperature change rate exceeds the secondary threshold (i.e., strain rate threshold and abrupt change rate threshold). If an abnormal strain rate (such as periodic deformation caused by vibration) or a sudden temperature change (such as local overheating) is detected, the high-speed acquisition mode is activated to capture full waveform data of the weld joint; if the triggering condition is not met, the data is buffered at a conventional frequency.

[0079] Generally, in-depth time-series analysis is required for the acquired full waveform or buffered data. Wavelet transform is used to extract the energy distribution of characteristic frequency bands in the light intensity signal, and fast Fourier transform is used to analyze the spectral characteristics of the vibration signal. This, combined with the time-varying pattern of the temperature gradient, helps identify fault modes. For example, when abnormal light intensity entropy values ​​are detected simultaneously in a specific frequency band, vibration energy is concentrated at a specific frequency, and the thermal hysteresis curve exhibits an opening characteristic, it can be identified as a "thermo-mechanical coupling fault."

[0080] Generally, labeled multi-dimensional parameters and corresponding fault types are used to form training samples for training a multimodal diagnostic model. This model employs a hybrid neural network architecture: the optical branch analyzes the temporal characteristics of light intensity sequences, the mechanical branch handles vibration spectrum modes, and the thermal branch learns the temperature field distribution. Finally, an adaptive weighted fusion module outputs a comprehensive diagnostic result. Through extensive sample training, the model can gradually establish an accurate mapping relationship from multi-dimensional parameters to fault types.

[0081] Optionally, based on the above embodiments, timing analysis processing is performed on each reference multi-dimensional parameter obtained from full waveform acquisition or short-time buffering to obtain the reference fault type corresponding to each reference multi-dimensional parameter, which may include:

[0082] The light intensity waveform signal in the optical feature parameters of the target reference multi-dimensional parameters is obtained and decomposed. The energy entropy of different feature frequency bands is calculated. When the entropy value of the detected energy entropy exceeds the preset threshold, the reference fault type corresponding to the target reference multi-dimensional parameters is determined to be a pollutant migration disturbance fault.

[0083] Vibration signals from mechanical characteristic parameters obtained from target reference multi-dimensional parameters are subjected to fast Fourier transform to calculate the energy proportion of different sensitive frequency bands. When an energy proportion value exceeds a preset threshold and is accompanied by abnormal light intensity signal, the reference fault type corresponding to the target reference multi-dimensional parameters is determined to be a resonance fault.

[0084] Thermal characteristic parameters are obtained from the target reference multidimensional parameters, and a response model of temperature gradient and light loss is established according to the thermodynamic coupling analysis algorithm. The thermal hysteresis curve corresponding to the response model is obtained through the cyclic temperature change test algorithm. When the area of ​​the open hysteresis loop in the thermal hysteresis curve is greater than the preset threshold, the reference fault type corresponding to the target reference multidimensional parameters is determined to be the thermal failure fault of the adhesive layer.

[0085] Generally, the analysis of optical characteristic parameters mainly involves multi-scale decomposition of the light intensity waveform signal. The light intensity signal is decomposed into different characteristic frequency bands using a wavelet packet transform algorithm, and the energy distribution entropy value of each frequency band is calculated. When the energy entropy of a specific frequency band exceeds a set threshold, it indicates that the signal in that frequency band exhibits abnormally disordered characteristics. This spectral characteristic is highly consistent with the light scattering phenomenon caused by the migration and disturbance of contaminants near the weld joint due to vibration; therefore, it can be identified as a contaminant migration disturbance fault.

[0086] In an optional implementation of this embodiment, at the optical signal analysis level, an improved wavelet packet transform is used to decompose the 10MHz (megahertz) sampled light intensity waveform into 8 layers to extract the energy entropy values ​​of three characteristic frequency bands: 120-150Hz (hertz), 150-300Hz, and 300-600Hz. When the entropy value exceeds 5.2, it is determined that there is a pollutant migration disturbance.

[0087] Generally, spectral analysis is used to process mechanical characteristic parameters. A Fast Fourier Transform is performed on the acquired vibration signal to calculate the proportion of energy in a preset sensitive frequency band relative to the total energy. When the proportion of energy in a specific frequency band is significantly higher than normal and is accompanied by abnormal light intensity, it indicates the presence of mechanical resonance. This simultaneous abnormality in mechanical vibration and optical parameters is consistent with the physical mechanism in resonance faults where mechanical energy affects light transmission characteristics through coupling at a specific frequency; therefore, it can be identified as a resonance fault.

[0088] In an optional implementation of this embodiment, the analysis employs resonant frequency matching technology to perform a 1024-point fast Fourier transform on the triaxial vibration signal, calculates the energy percentage in the sensitive frequency band of 1.2-3.5 kHz (kilohertz), and triggers a resonance fault alarm when the energy percentage exceeds 65% and is accompanied by abnormal light intensity.

[0089] Generally, the analysis of thermal characteristic parameters requires the establishment of a thermo-optical coupling model. A thermodynamic coupling analysis algorithm is used to construct the correspondence between temperature gradient and optical loss, and then cyclic temperature change tests are conducted to obtain the thermal hysteresis curve. When the curve exhibits a clear opening shape and the opening area exceeds a threshold, it indicates that the material exhibits irreversible heat dissipation characteristics during heating and cooling. This thermal hysteresis phenomenon is consistent with the typical characteristics of aging failure of adhesive materials under alternating temperature conditions; therefore, it can be identified as a thermal failure of the adhesive layer.

[0090] In an optional implementation of this embodiment, a response model of temperature gradient and light loss is established by thermodynamic coupling analysis. The thermal hysteresis curve is obtained by cyclic temperature change test from -40℃ to 85℃. When the area of ​​the open hysteresis loop is greater than 3dB℃, the adhesive layer is determined to be thermally failed.

[0091] S260. After generating single-dimensional identification results corresponding to optical, mechanical and thermal characteristic parameters respectively through the multimodal diagnostic model, the single-dimensional identification results are fused to obtain and output the final fused fault type and fused fault confidence.

[0092] S270. When the confidence level of the fusion failure is greater than or equal to the confidence threshold, trigger a multiphysics stress loading test on the key fusion joint and check whether the fusion failure type can be reproduced during the test.

[0093] S280. If a fault of the fusion fault type can be reproduced during the test, the fusion fault type is output as the intermittent fault diagnosis result of the key fusion point.

[0094] The technical solution of this invention collects light intensity changes by deploying fiber optic couplers and photodetectors upstream and downstream of key fusion splices, simultaneously using fiber optic circulators and optical power meters to collect back reflectivity, and then collecting axial strain values ​​through fiber optic grating sensors arranged at preset intervals. Temperature distribution data and temperature gradient data are also collected through distributed temperature-measuring fibers wound around the surface of the protective sleeve. These multi-dimensional parameters are input into a pre-trained multi-modal diagnostic model. This model first generates single-dimensional identification results corresponding to optical, mechanical, and thermal characteristic parameters, respectively. Then, it fuses these single-dimensional identification results to obtain and output the final fused fault type. The method integrates the fusion fault confidence level. When the fusion fault confidence level is greater than or equal to the confidence level threshold, a multi-physics stress loading test is triggered on the key fusion splice. The test checks whether the fusion fault type can be reproduced during the test. If reproduction is detected, the fusion fault type is output as the intermittent fault diagnosis result for the key fusion splice. This novel intermittent fault identification method for fiber optic gyroscope fusion splices achieves comprehensive real-time perception of optical transmission performance, mechanical deformation state, and thermal environment parameters through multi-sensor collaboration, providing a rich data foundation for fault diagnosis. At the same time, by combining intelligent diagnosis with physical verification, the accuracy and reliability of intermittent fault identification are significantly improved.

[0095] Example 3

[0096] Figure 3This is a flowchart of another intermittent fault identification method for fiber optic gyroscope fusion splices provided in Embodiment 2 of the present invention. This embodiment is based on the above embodiments and optimized. Specifically, the operation of "after generating single-dimensional identification results corresponding to optical feature parameters, mechanical feature parameters and thermal feature parameters respectively through the multi-modal diagnostic model, fusing the single-dimensional identification results to obtain and output the final fused fault type and fused fault confidence" has been refined.

[0097] Correspondingly, such as Figure 3 As shown, the method includes:

[0098] S310. By deploying multiple types of sensors at the key fusion splice points of the fiber optic gyroscope, multi-dimensional parameters of the key fusion splice points are collected synchronously; among which, the multi-dimensional parameters include optical characteristic parameters, mechanical characteristic parameters and thermal characteristic parameters.

[0099] S320. Input the multi-dimensional parameters of the key weld points into the pre-trained multimodal diagnostic model.

[0100] S330. The multimodal diagnostic model converts the identification results of each dimension into corresponding confidence vectors.

[0101] In this embodiment, the model's preliminary identification conclusions for parameters of each dimension are transformed into a quantifiable probabilistic expression. By numerically encapsulating the judgment results of different attributes such as optical anomalies, mechanical vibration characteristics, and temperature field changes, a vector structure with unified dimensions is formed to facilitate subsequent fusion calculations.

[0102] S340. Input each confidence vector into the attention mechanism module in the multimodal diagnostic model, and calculate and assign weight coefficients for fusion based on the uncertainty measure of the identification results of each dimension.

[0103] In this embodiment, a specific weighting mechanism is used to address the differences in the reliability of recognition results across different dimensions. Based on the inherent characteristics such as optical signal stability, mechanical vibration data integrity, and temperature monitoring accuracy, the contribution ratio of each dimension vector in the final decision is dynamically adjusted, so that data with higher reliability receives greater decision weight.

[0104] S350. By using a fusion algorithm, the weighted confidence vectors of each dimension are fused together, and the confidence intervals of all candidate fused fault types are calculated.

[0105] In this embodiment, an evidence theory synthesis algorithm is used to integrate and calculate the weighted multidimensional vectors. By establishing a hypothesis space and a confidence assignment function, the confidence range of various potential failure modes is derived, thereby obtaining the confidence interval estimation result for each possible failure type.

[0106] S360. Based on the reliability interval results of all candidate fault types, select the fusion candidate fault type with the highest upper limit of the reliability interval as the final determined fusion fault type, and use the upper limit of the reliability interval as the confidence level of the fusion fault in this diagnosis.

[0107] In this embodiment, decision-making is based on the confidence interval data obtained from fusion calculation. By comparing the upper limits of the confidence intervals of different fault hypotheses, the fault type with the highest confidence value is selected as the final conclusion, and this confidence value is used as the deterministic measure of the diagnosis result.

[0108] S370. When the confidence level of the fusion failure is greater than or equal to the confidence threshold, trigger a multiphysics stress loading test on the key fusion joint and check whether the fusion failure type can be reproduced during the test.

[0109] Optionally, based on the above embodiments, triggering a multiphysics stress loading test on the critical weld joint may include:

[0110] The integrated control platform coordinates the start-up of the three-axis hydraulic vibration table and the programmable temperature and humidity chamber according to the preset program.

[0111] The triaxial hydraulic vibration table is controlled to apply a wideband random vibration spectrum covering the low to high frequency range, while the programmable temperature and humidity chamber is controlled to perform a rapid temperature cycle change including extreme low temperature and extreme high temperature.

[0112] By deploying multiple types of sensors at the key weld points, the transient response of optical power, dynamic strain distribution, and temperature field changes of the weld points under composite stress environment are collected synchronously and continuously.

[0113] The system compares the collected parameter change curves with the characteristic patterns of the target fusion fault type in real time. When a highly matching fault characteristic signal is detected, it immediately locks and records the vibration spectrum parameters and temperature cycle conditions that trigger the characteristic.

[0114] Generally, an integrated control platform is used to coordinate the control of the triaxial hydraulic vibration table and the programmable temperature and humidity chamber, ensuring that the two stress loading devices can start and run synchronously according to a preset program, thereby constructing a multi-physics coupled testing environment. This coordinated control mechanism ensures the precise matching of vibration stress and temperature stress over time, providing a reliable stress condition basis for subsequent fault reproduction.

[0115] Generally, during stress loading, the vibration table outputs a wide-band random vibration spectrum covering a range from low to high frequencies, simulating the complex mechanical vibration environment in actual working conditions. Simultaneously, the temperature and humidity chamber performs rapid temperature cycling changes, including extreme low and high temperatures, to reproduce the impact of temperature shocks on weld joints. The simultaneous application of these two stresses can effectively induce intermittent fault characteristics that only appear under specific combined stress conditions.

[0116] Generally, during the application of combined stress, multiple sensors deployed around the weld joint simultaneously collect multi-dimensional parameters such as transient response of optical power, dynamic strain distribution, and temperature field changes. These parameters reflect real-time data on the optical transmission characteristics, mechanical deformation state, and thermal environment changes of the weld joint under the combined effects of mechanical vibration and temperature shock, providing comprehensive data support for fault feature identification.

[0117] Generally, by comparing the collected parameter change curves with the characteristic patterns of the target fault type in real time, when key indicators such as light intensity fluctuation patterns and spectral characteristics are highly matched with the target fault characteristics, the specific vibration spectrum parameters and temperature cycle conditions corresponding to triggering that characteristic are immediately recorded. For example, when the characteristic of periodic attenuation of optical power under the combined effect of a specific frequency vibration and rapid temperature change is detected, it can be confirmed that the fault mode has been successfully reproduced under the current stress conditions.

[0118] S380. If a fault of the fusion fault type can be reproduced during the test, the fusion fault type is output as the intermittent fault diagnosis result of the key fusion point.

[0119] Optionally, based on the above embodiments, detecting whether the fusion failure type can be reproduced during the test may include:

[0120] By synchronously collecting parameters from various dimensions during the multi-physics stress loading test, temporal features that can characterize the dynamic evolution of the fault are extracted, including but not limited to the transient fluctuation pattern of optical power, the abrupt change period of reflectivity, and the time-varying law of polarization crosstalk.

[0121] Using a pre-defined similarity measurement algorithm, the degree of matching between the extracted temporal features and the standard feature template corresponding to the target fused fault type is calculated to obtain a quantitative similarity value.

[0122] The calculated similarity value is compared with the preset judgment threshold, and at the same time it is checked whether the currently observed fault duration matches the typical duration recorded in historical cases. The final judgment of successful fault reproduction is made if and only if the two conditions of similarity exceeding the threshold and duration matching are met simultaneously.

[0123] Record the specific vibration spectrum parameters, temperature cycling conditions, and the complete set of observed fault characteristic parameters applied when the fault reproduction is achieved, and form a verification report containing stress conditions and response characteristics as substantial evidence to support the diagnostic conclusion.

[0124] Generally, multi-physics stress loading tests are used to actively induce and capture fault characteristics. During the test, multi-dimensional parameters such as optical, mechanical and thermal parameters are collected simultaneously, and key temporal features that can reflect the dynamic evolution of the fault are extracted from them, such as the transient fluctuation pattern of optical power, the abrupt change periodic law of reflectivity and the time-varying characteristics of polarization crosstalk. These features together constitute the dynamic fingerprint of fault identification.

[0125] Generally, after obtaining time-series features, a specific similarity measurement algorithm is needed to quantify the degree of matching between these features and the standard template of the target fault type. This algorithm comprehensively considers multiple dimensions such as waveform similarity, periodic consistency, and overlap of change trends, and finally outputs a quantified similarity value to provide an objective basis for subsequent judgment.

[0126] Generally, the final determination of fault reproduction requires the simultaneous fulfillment of two key conditions: first, the feature similarity value must exceed a preset threshold, indicating that the current phenomenon closely matches the target fault; second, the observed fault duration must match the typical duration recorded in historical cases, ensuring that this is not an accidental transient phenomenon. Only when both conditions are met can a reliable conclusion of successful fault reproduction be drawn.

[0127] Generally, after confirming the reproduction of a fault, it is necessary to record three key pieces of information: first, the specific vibration spectrum parameters applied when the fault was triggered (such as frequency range and vibration intensity); second, the detailed conditions of the temperature cycle (such as temperature range and rate of change); and third, the complete set of observed fault characteristic parameters. This information will be integrated to form a complete verification report, providing solid experimental evidence to support the diagnostic conclusions.

[0128] The technical solution of this invention involves simultaneously collecting multi-dimensional parameters, including optical, mechanical, and thermal features, from multiple sensors deployed at key fusion splices of a fiber optic gyroscope. These multi-dimensional parameters are then input into a pre-trained multimodal diagnostic model. The model converts the identification results of each dimension into corresponding confidence vectors. Its internal attention mechanism module calculates and assigns fusion weight coefficients based on the uncertainty measure of each dimension's identification results. A fusion algorithm then fuses the weighted confidence vectors of each dimension to calculate the confidence intervals for all candidate fault types. The candidate fault type with the highest upper limit of the confidence interval is selected as the final determined fusion fault type, and its upper limit is used as the fusion fault confidence. When this confidence is greater than or equal to the confidence threshold, a multi-physics stress loading test is triggered at the key fusion splice to check if the fault can be reproduced. If reproduction is successful, the fusion fault type is output as the diagnostic result. This novel intermittent fault identification method for fiber optic gyroscope fusion splices effectively improves the decision reliability of multi-source information fusion through confidence vector transformation and dynamic weight fusion mechanisms. Simultaneously, it significantly enhances the identification accuracy of complex fault modes through a confidence interval optimization strategy.

[0129] Example 4

[0130] Figure 4 This is a schematic diagram of an intermittent fault identification device for a fiber optic gyroscope fusion splice provided in Embodiment 4 of the present invention. Figure 4 As shown, the device includes: a synchronous data acquisition module 410, a data input module 420, a multimodal fusion decision module 430, a detection and reproduction module 440, and a fusion output module 450, wherein:

[0131] The synchronous data acquisition module 410 is used to synchronously acquire multi-dimensional parameters of the key fusion splice point through multiple types of sensors deployed at the key fusion splice point of the fiber optic gyroscope; wherein, the multi-dimensional parameters include optical characteristic parameters, mechanical characteristic parameters and thermal characteristic parameters;

[0132] Data input module 420 is used to input the multi-dimensional parameters of the key weld points into a pre-trained multimodal diagnostic model;

[0133] The multimodal fusion decision module 430 is used to fuse the single-dimensional identification results corresponding to optical feature parameters, mechanical feature parameters and thermal feature parameters respectively through the multimodal diagnostic model, and obtain and output the final fused fault type and fused fault confidence.

[0134] The detection and reproduction module 440 is used to trigger a multiphysics stress loading test on the key weld joint when the confidence level of the fusion failure is greater than or equal to the confidence level threshold, and to detect whether the fusion failure type of failure can be reproduced during the test.

[0135] The fusion output module 450 is used to output the fusion fault type as an intermittent fault diagnosis result for the key weld point when a fault of the fusion fault type that can be reproduced during the test is detected.

[0136] The technical solution of this invention involves simultaneously acquiring multi-dimensional parameters, including optical, mechanical, and thermal characteristic parameters, from multiple sensors deployed at key fusion splices of a fiber optic gyroscope. These multi-dimensional parameters are then input into a pre-trained multimodal diagnostic model. The multimodal diagnostic model first generates single-dimensional identification results corresponding to the optical, mechanical, and thermal characteristic parameters, respectively. Subsequently, the single-dimensional identification results are fused to obtain and output the final fused fault type and fused fault confidence level. When the fused fault confidence level is greater than or equal to a confidence threshold, a multi-physics stress loading test is triggered on the key fusion splice, and it is checked whether the fused fault type can be reproduced during the test. If reproduction is successful, the fused fault type is output as the intermittent fault diagnosis result for the key fusion splice. This novel intermittent fault identification method for fiber optic gyroscope fusion splices effectively solves the technical bottleneck of traditional detection methods in capturing transient anomalies and reproducing sporadic faults. Through the mechanism of multi-source information fusion and active stress loading verification, the accuracy and reliability of fault diagnosis are significantly improved.

[0137] Based on the above embodiments, the optical characteristic parameters include light intensity change and back reflectivity; the mechanical characteristic parameters include axial strain value; and the thermal characteristic parameters include temperature distribution data and temperature gradient data.

[0138] Based on the above embodiments, the synchronous data acquisition module 410 can be specifically used for:

[0139] The light intensity change at the critical fusion splice is collected by fiber optic couplers and photodetectors set up upstream and downstream of the critical fusion splice.

[0140] The back reflectivity of the critical fusion splice was collected by using fiber optic circulators and optical power meters placed upstream and downstream of the critical fusion splice.

[0141] The axial strain value of the critical fusion splice is collected by fiber optic grating sensors arranged at preset intervals on both sides of the critical fusion splice.

[0142] Temperature distribution and temperature gradient data of key fusion splices are collected by distributed temperature-measuring optical fibers wound around the surface of the protective sleeve.

[0143] Furthermore, based on the above embodiments, the intermittent fault identification device for fiber optic gyroscope splices may further include:

[0144] The data acquisition module is used to collect the light intensity change and back reflectivity of multiple reference fusion points on multiple reference fiber optic gyroscopes in real time before inputting the multi-dimensional parameters of the key fusion points into the pre-trained multimodal diagnostic model.

[0145] The multi-level trigger acquisition module is used to continue acquiring axial strain values ​​and temperature distribution data on the target reference weld point if the light intensity decrease rate of the target reference weld point exceeds a preset decrease rate threshold within a preset time period, or the reflectivity increase rate exceeds a preset increase rate threshold within a preset time period.

[0146] The adaptive sampling control module is used to collect the full waveform of the reference multi-dimensional parameters of the target reference weld point at a preset sampling frequency if the mechanical strain rate of the target reference weld point exceeds a preset strain rate threshold within a preset time period, or the temperature change rate exceeds a preset change rate threshold within a preset time period; otherwise, it performs short-term buffering of the reference multi-dimensional parameters of the target reference weld point.

[0147] The timing feature analysis module is used to perform timing analysis processing based on the reference multi-dimensional parameters acquired from full waveform acquisition or short-time buffering, and to obtain the reference fault type corresponding to each reference multi-dimensional parameter.

[0148] The model training module is used to construct multiple training samples using each reference multi-dimensional parameter and the reference fault type corresponding to each reference multi-dimensional parameter, and to train the preset machine learning model using each training sample to obtain a multimodal diagnostic model.

[0149] Based on the above embodiments, the time series feature analysis module can be specifically used for:

[0150] The light intensity waveform signal in the optical feature parameters of the target reference multi-dimensional parameters is obtained and decomposed. The energy entropy of different feature frequency bands is calculated. When the entropy value of the detected energy entropy exceeds the preset threshold, the reference fault type corresponding to the target reference multi-dimensional parameters is determined to be a pollutant migration disturbance fault.

[0151] Vibration signals from mechanical characteristic parameters obtained from target reference multi-dimensional parameters are subjected to fast Fourier transform to calculate the energy proportion of different sensitive frequency bands. When an energy proportion value exceeds a preset threshold and is accompanied by abnormal light intensity signal, the reference fault type corresponding to the target reference multi-dimensional parameters is determined to be a resonance fault.

[0152] Thermal characteristic parameters are obtained from the target reference multidimensional parameters, and a response model of temperature gradient and light loss is established according to the thermodynamic coupling analysis algorithm. The thermal hysteresis curve corresponding to the response model is obtained through the cyclic temperature change test algorithm. When the area of ​​the open hysteresis loop in the thermal hysteresis curve is greater than the preset threshold, the reference fault type corresponding to the target reference multidimensional parameters is determined to be the thermal failure fault of the adhesive layer.

[0153] Based on the above embodiments, the multimodal fusion decision module 430 can be specifically used for:

[0154] The multimodal diagnostic model converts the identification results of each dimension into corresponding confidence vectors.

[0155] Each confidence vector is input into the attention mechanism module in the multimodal diagnostic model, and weight coefficients for fusion are calculated and assigned based on the uncertainty measure of the identification results of each dimension.

[0156] By using a fusion algorithm, the weighted confidence vectors of each dimension are fused together, and the confidence intervals of all candidate fused fault types are calculated.

[0157] Based on the reliability interval results of all candidate fault types, the fusion candidate fault type with the highest upper limit of the reliability interval is selected as the final determined fusion fault type, and the upper limit of the reliability interval is used as the confidence level of the fusion fault in this diagnosis.

[0158] Based on the above embodiments, the detection and reproduction module 440 can be specifically used for:

[0159] The integrated control platform coordinates the start-up of the three-axis hydraulic vibration table and the programmable temperature and humidity chamber according to the preset program.

[0160] The triaxial hydraulic vibration table is controlled to apply a wideband random vibration spectrum covering the low to high frequency range, while the programmable temperature and humidity chamber is controlled to perform a rapid temperature cycle change including extreme low temperature and extreme high temperature.

[0161] By deploying multiple types of sensors at the key weld points, the transient response of optical power, dynamic strain distribution, and temperature field changes of the weld points under composite stress environment are collected synchronously and continuously.

[0162] The system compares the collected parameter change curves with the characteristic patterns of the target fusion fault type in real time. When a highly matching fault characteristic signal is detected, it immediately locks and records the vibration spectrum parameters and temperature cycle conditions that trigger the characteristic.

[0163] Based on the above embodiments, the fusion output module 450 can be specifically used for:

[0164] By synchronously collecting parameters from various dimensions during the multi-physics stress loading test, temporal features that can characterize the dynamic evolution of the fault are extracted, including but not limited to the transient fluctuation pattern of optical power, the abrupt change period of reflectivity, and the time-varying law of polarization crosstalk.

[0165] Using a pre-defined similarity measurement algorithm, the degree of matching between the extracted temporal features and the standard feature template corresponding to the target fused fault type is calculated to obtain a quantitative similarity value.

[0166] The calculated similarity value is compared with the preset judgment threshold, and at the same time it is checked whether the currently observed fault duration matches the typical duration recorded in historical cases. The final judgment of successful fault reproduction is made if and only if the two conditions of similarity exceeding the threshold and duration matching are met simultaneously.

[0167] Record the specific vibration spectrum parameters, temperature cycling conditions, and the complete set of observed fault characteristic parameters applied when the fault reproduction is achieved, and form a verification report containing stress conditions and response characteristics as substantial evidence to support the diagnostic conclusion.

[0168] The intermittent fault identification device for fiber optic gyroscope fusion splices provided in this embodiment of the invention can execute the intermittent fault identification method for fiber optic gyroscope fusion splices provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0169] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0170] Example 5

[0171] Figure 5 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0172] like Figure 5As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0173] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0174] Processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as performing an intermittent fault identification method for fiber optic gyroscope splices as described in any embodiment of the present invention, namely:

[0175] Multiple types of sensors are deployed at the key fusion splice points of the fiber optic gyroscope to synchronously collect multi-dimensional parameters of the key fusion splice points; among these, the multi-dimensional parameters include optical characteristic parameters, mechanical characteristic parameters, and thermal characteristic parameters.

[0176] The multi-dimensional parameters of the key weld points are input into a pre-trained multimodal diagnostic model;

[0177] After generating single-dimensional identification results corresponding to optical, mechanical, and thermal characteristic parameters respectively through the multimodal diagnostic model, the single-dimensional identification results are fused to obtain and output the final fused fault type and fused fault confidence.

[0178] When the confidence level of the fusion failure is greater than or equal to the confidence threshold, a multiphysics stress loading test is triggered on the key fusion joint, and it is checked whether the fusion failure type can be reproduced during the test.

[0179] If so, the fusion fault type is output as the intermittent fault diagnosis result for the critical fusion point.

[0180] In some embodiments, a method for identifying intermittent faults in a fiber optic gyroscope fusion splice as described in any one of the embodiments of the present invention can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the method for identifying intermittent faults in a fiber optic gyroscope fusion splice as described above can be performed. Alternatively, in other embodiments, processor 11 can be configured by any other suitable means (e.g., by means of firmware) to perform the method for identifying intermittent faults in a fiber optic gyroscope fusion splice as described in any one of the embodiments of the present invention.

[0181] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0182] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0183] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0184] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0185] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0186] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0187] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0188] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for identifying intermittent faults of a fiber-optic gyroscope fusion splice point, characterized by, The method comprises: Synchronously collecting multi-dimensional parameters of the key fusion point through a plurality of sensors deployed at the key fusion point of the fiber optic gyroscope, wherein the multi-dimensional parameters include optical characteristic parameters, mechanical characteristic parameters and thermal characteristic parameters; Inputting the multi-dimensional parameters of the key fusion point into a pre-trained multi-modal diagnostic model; After generating single-dimensional recognition results corresponding to the optical characteristic parameters, the mechanical characteristic parameters and the thermal characteristic parameters respectively through the multi-modal diagnostic model, fusing the single-dimensional recognition results to obtain and output a final fusion fault type and a fusion fault confidence; When the fusion fault confidence is greater than or equal to a confidence threshold, triggering a multi-physical field stress loading test of the key fusion point, and detecting whether the fusion fault type can be reproduced in the test process; If yes, outputting the fusion fault type as an intermittent fault diagnosis result of the key fusion point; The triggering of the multi-physical field stress loading test of the key fusion point comprises: Starting a three-axis hydraulic vibration table and a programmable temperature and humidity box according to a preset program through an integrated control platform; controlling the three-axis hydraulic vibration table to apply a wideband random vibration spectrum covering a low frequency to a high frequency range, and controlling the programmable temperature and humidity box to perform a rapid temperature cycle change including an extreme low temperature and an extreme high temperature; synchronously and continuously collecting optical power transient response, dynamic strain distribution and temperature field change data of the fusion point under a combined stress environment through a plurality of sensors deployed at the key fusion point; comparing the collected parameter change curves with a characteristic mode of a target fusion fault type in real time, and when a highly matched fault characteristic signal is monitored, locking and recording vibration spectrum parameters and temperature cycle conditions corresponding to the triggering of the characteristic; The detection of whether the fusion fault type can be reproduced in the test process comprises: Extracting time sequence features capable of representing a dynamic evolution process of the fault from the synchronously collected multi-dimensional parameters in the multi-physical field stress loading test process, wherein the time sequence features include but are not limited to transient fluctuation patterns of optical power, mutation periods of reflectivity and time-varying laws of polarization crosstalk; calculating a matching degree between the extracted time sequence features and a standard feature template corresponding to the target fusion fault type by using a preset similarity measurement algorithm to obtain a quantitative similarity value; comparing the calculated similarity value with a preset determination threshold, and simultaneously verifying whether a current observed fault duration is consistent with a typical duration recorded in a historical case; when and only when both the similarity is over the threshold and the duration is matched, a final determination of successful reproduction of the fault is made; recording specific vibration spectrum parameters, temperature cycle conditions and a complete set of observed fault characteristic parameters applied when the fault is reproduced to form a verification report containing stress conditions and response features as substantial evidence supporting the diagnostic conclusion.

2. The method of claim 1, wherein, The optical characteristic parameters include light intensity variation and back reflectivity; the mechanical characteristic parameters include axial strain values; and the thermal characteristic parameters include temperature distribution data and temperature gradient data. Correspondingly, by deploying multiple types of sensors at the key fusion points of the fiber optic gyroscope, the multi-dimensional parameters of the key fusion points are synchronously collected, including: By setting fiber couplers and photodetectors upstream and downstream of the key fusion points, the light intensity variation of the key fusion points is collected; By setting fiber optic circulators and optical power meters upstream and downstream of the key fusion points, the back reflectivity of the key fusion points is collected; By deploying fiber Bragg grating sensors at a preset interval on both sides of the key fusion points, the axial strain value of the key fusion points is collected; By deploying distributed temperature measurement fibers on the surface of the key fusion point protection sleeve, the temperature distribution data and temperature gradient data of the key fusion points are collected.

3. The method of claim 2, wherein, Before inputting the multi-dimensional parameters of the key fusion points into the pre-trained multi-modal diagnostic model, the method further includes: Real-time collection of light intensity variation and back reflectivity of multiple reference fusion points on multiple reference fiber optic gyroscopes; If the light intensity drop rate of the target reference fusion point within a preset time period exceeds a preset drop rate threshold, or the reflectivity surge rate within a preset time period exceeds a preset surge rate threshold, then continue to collect the axial strain value and temperature distribution data on the target reference fusion point; If the mechanical strain rate of the target reference fusion point within a preset time period exceeds a preset strain rate threshold, or the temperature sudden change rate within a preset time period exceeds a preset sudden change rate threshold, then perform full waveform collection of the reference multi-dimensional parameters of the target reference fusion point at a preset collection frequency, otherwise, perform short-time caching of the reference multi-dimensional parameters of the target reference fusion point; Time series analysis and processing of the full waveform collection or short-time caching of each reference multi-dimensional parameter to obtain a reference fault type corresponding to each reference multi-dimensional parameter; Using each reference multi-dimensional parameter and the reference fault type corresponding to each reference multi-dimensional parameter to construct multiple training samples, and using each training sample to train a preset machine learning model to obtain a multi-modal diagnostic model.

4. The method of claim 3, wherein, Time series analysis and processing of the full waveform collection or short-time caching of each reference multi-dimensional parameter to obtain a reference fault type corresponding to each reference multi-dimensional parameter, including: Decompose the light intensity waveform signal in the optical characteristic parameter in the target reference multi-dimensional parameter, calculate the energy entropy of different characteristic frequency bands, and when an energy entropy value exceeding a preset threshold is detected, determine that the reference fault type corresponding to the target reference multi-dimensional parameter is a contaminant migration disturbance fault; Perform fast Fourier transform on the vibration signal in the mechanical characteristic parameter in the target reference multi-dimensional parameter, calculate the energy proportion of different sensitive frequency bands, and when an energy proportion value exceeding a preset threshold is detected and accompanied by an abnormal light intensity signal, determine that the reference fault type corresponding to the target reference multi-dimensional parameter is a resonance fault; The thermal characteristic parameter is obtained in the target reference multidimensional parameter, and a response model of temperature gradient and optical loss is established according to a thermodynamic coupling analysis algorithm; a thermal hysteresis curve corresponding to the response model is obtained through a cyclic temperature variation test algorithm, and when an opening type hysteresis loop area in the thermal hysteresis curve is greater than a preset threshold, it is determined that the reference fault type corresponding to the target reference multidimensional parameter is a fixed adhesive layer thermal failure fault.

5. The method according to any one of claims 1 to 4, characterized in that, After the multi-modal diagnosis model generates single-dimensional identification results corresponding to the optical characteristic parameter, the mechanical characteristic parameter and the thermal characteristic parameter respectively, the single-dimensional identification results are fused to obtain and output a final fusion fault type and a fusion fault confidence, including: The multi-modal diagnosis model converts each dimensional identification result into a corresponding confidence vector; The attention mechanism module in the multi-modal diagnosis model is inputted with the confidence vectors, and a weight coefficient for fusion is calculated and distributed according to the uncertainty measurement of each dimensional identification result itself; Through a fusion algorithm, the weighted dimensional confidence vectors are fused to calculate the confidence interval of all candidate fusion fault types; According to the confidence interval results of all candidate fault types, the fusion candidate fault type with the highest upper limit of the confidence interval is selected as the final determined fusion fault type, and the upper limit of the confidence interval is taken as the fusion fault confidence of this diagnosis.

6. An intermittent fault recognition device for a fiber-optic gyroscope fusion splice point, characterized by, The device comprises: A synchronous data acquisition module is configured to acquire multidimensional parameters of a key fusion point of an FOG through a plurality of sensors arranged at the key fusion point, wherein the multidimensional parameters include optical characteristic parameters, mechanical characteristic parameters and thermal characteristic parameters; A data input module is configured to input the multidimensional parameters of the key fusion point into a pre-trained multi-modal diagnosis model; A multi-modal fusion decision module is configured to fuse single-dimensional identification results corresponding to the optical characteristic parameters, the mechanical characteristic parameters and the thermal characteristic parameters respectively through the multi-modal diagnosis model to obtain and output a final fusion fault type and a fusion fault confidence; A detection reproduction module is configured to trigger a multi-physical field stress loading test of the key fusion point when the fusion fault confidence is greater than or equal to a confidence threshold, and detect whether the fusion fault type can be reproduced in the test process; A fusion output module is configured to output the fusion fault type as an intermittent fault diagnosis result of the key fusion point when it is detected that the fusion fault type can be reproduced in the test process. The detection reproduction module is specifically configured to: drive the three-axis hydraulic vibration table and the programmable temperature and humidity box to start according to a preset program through the integrated control platform; control the three-axis hydraulic vibration table to apply a wideband random vibration spectrum covering a low frequency to a high frequency range, and control the programmable temperature and humidity box to execute a rapid temperature cycle change including an extreme low temperature and an extreme high temperature; synchronously and continuously collect light power transient response, dynamic strain distribution and temperature field change data of the key fusion point in a composite stress environment through the multiple types of sensors arranged at the key fusion point; compare the collected parameter change curves with a characteristic mode of a target fusion failure type in real time, and when a highly matched failure characteristic signal is monitored, the vibration spectrum parameters and the temperature cycle conditions corresponding to the characteristic are immediately locked and recorded; extract time sequence characteristics capable of representing a dynamic evolution process of the failure from the synchronously collected parameters in each dimension during the multi-physical field stress loading test, wherein the time sequence characteristics include but are not limited to transient fluctuation patterns of light power, mutation periods of reflectivity and time-varying laws of polarization crosstalk; calculate a matching degree between the extracted time sequence characteristics and a standard characteristic template corresponding to the target fusion failure type by using a preset similarity measurement algorithm, and obtain a quantitative similarity value; compare the calculated similarity value with a preset determination threshold, and simultaneously verify whether a current observed failure duration matches a typical duration recorded in a historical case, and when and only when both the similarity exceeds the threshold and the duration matches, a final determination of successfully reproducing the failure is made; record specific vibration spectrum parameters, temperature cycle conditions and a complete set of observed failure characteristic parameters applied when the failure reproduction is achieved, and form a verification report including stress conditions and response characteristics as substantial evidence supporting a diagnosis conclusion.

7. An electronic device, comprising: The electronic device includes: at least one processor; and a memory connected to the at least one processor in communication; wherein The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the intermittent fault identification method of the fiber-optic gyroscope fusion point according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing the processor to execute the intermittent fault identification method of the fiber-optic gyroscope fusion point according to any one of claims 1-5.

9. A computer program product, characterised in that, The computer program product includes a computer program that, when executed by a processor, implements the intermittent fault identification method of the fiber-optic gyroscope fusion point according to any one of claims 1-5.

Images