Optoelectronic device fault detection platform based on multi-source data fusion
Through the multi-source data fusion platform, the multi-modal monitoring data of optoelectronic devices is collected, preprocessed and evaluated, and the problems of low fault detection efficiency and low accuracy in the existing technology are solved, and more accurate and efficient fault detection is achieved, improving the The reliability of device operation and system stability are achieved.
Patent Information
- Application Number
- CN202411037900.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2024-11-19
AI Technical Summary
Existing optoelectronic device fault detection methods are inefficient and have low accuracy, especially in complex working conditions, it is difficult to effectively integrate multi-source heterogeneous data, which affects the comprehensiveness and accuracy of fault detection.
Adopting an optoelectronic device fault detection platform based on multi-source data fusion, multi-modal monitoring data is collected through multi-source data acquisition unit, data preprocessing unit performs denoising and feature extraction, health status evaluation unit performs device health status evaluation, and fault diagnosis unit Determine the fault status based on the evaluation results and build a full-process fault detection framework.
It improves the comprehensiveness, accuracy and efficiency of fault detection of optoelectronic devices, promptly discovers device performance degradation trends, generates accurate fault diagnosis results, and improves the reliability of the system and long-term fault-free operation capabilities.
Smart Images

Figure CN118981678A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of optoelectronic device detection, and in particular to an optoelectronic device fault detection platform based on multi-source data fusion. Background Art
[0002] Optoelectronic devices are affected by temperature, humidity, vibration, electromagnetic interference and other factors in the actual working environment, which may lead to the degradation of device performance or even failure, affecting the reliability and stability of the system. Therefore, it is very important to effectively detect faults in optoelectronic devices. Existing methods for fault detection of optoelectronic devices mainly include rule-based threshold judgment method and data-driven machine learning method. The rule-based threshold judgment method sets the threshold of each monitoring parameter. When the monitoring data exceeds the threshold, it is judged as a fault. However, this method is difficult to cope with complex working conditions, and the setting of the threshold depends on expert experience and is highly subjective. The data-driven machine learning method automatically learns fault features and identifies fault types by constructing a machine learning model, but the existing research lacks the fusion analysis of multi-source heterogeneous data, which affects the accuracy and comprehensiveness of fault detection. Therefore, the efficiency and accuracy of fault detection of optoelectronic devices in the prior art are low. Summary of the invention
[0003] The present application aims to solve the technical problems of low efficiency and low accuracy of optoelectronic device fault detection in the prior art by providing an optoelectronic device fault detection platform based on multi-source data fusion.
[0004] The optoelectronic device fault detection platform based on multi-source data fusion disclosed in the present application includes: a multi-source data acquisition unit, which is used to acquire multi-source monitoring data of the operating environment of the target optoelectronic device, and the multi-source monitoring data includes multiple parameter monitoring results reflecting the working state of the target optoelectronic device; a data preprocessing unit, which is used to preprocess the multi-source monitoring data and generate a monitoring data set for device health status assessment; a health status assessment unit, which is used to assess the health status of the target optoelectronic device based on the monitoring data set and obtain a health status assessment result; a fault diagnosis unit, which is used to determine whether the target optoelectronic device is in a fault state based on the health status assessment result, and generate a fault diagnosis result when the optoelectronic device is in a fault state.
[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0006] The multi-source monitoring data of the operating environment of the target optoelectronic device is collected by a multi-source data collection unit, including multiple parameter monitoring results reflecting the working state of the target optoelectronic device, which comprehensively characterizes the working state of the optoelectronic device and provides an information basis for subsequent fault detection; the collected multi-source monitoring data is preprocessed by a data preprocessing unit to generate a monitoring data set for device health status assessment, improve data quality, reduce the impact of data redundancy and noise, and generate a monitoring data set to facilitate subsequent health status assessment and fault diagnosis; the health status assessment unit assesses the health status of the target optoelectronic device based on the generated monitoring data set, obtains the health status assessment result, and promptly discovers the performance degradation trend of the device; the fault diagnosis unit determines whether the target optoelectronic device is in a fault state based on the health status assessment result; when the optoelectronic device is determined to be in a fault state, a technical solution of generating a fault diagnosis result solves the technical problems of low efficiency and low accuracy in fault detection of optoelectronic devices in the prior art, and achieves the technical effect of improving the efficiency and accuracy of fault detection of optoelectronic devices.
[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Figure 1 A structural schematic diagram of an optoelectronic device fault detection platform based on multi-source data fusion is provided for an embodiment of the present application;
[0009] Figure 2 A flow chart of obtaining a health status assessment model in an optoelectronic device fault detection platform based on multi-source data fusion is provided for an embodiment of the present application.
[0010] Explanation of the reference numerals: multi-source data acquisition unit 11 , data preprocessing unit 12 , health status assessment unit 13 , fault diagnosis unit 14 . DETAILED DESCRIPTION
[0011] The overall idea of the technical solution provided by this application is as follows:
[0012] An embodiment of the present application provides an optoelectronic device fault detection platform based on multi-source data fusion. First, a multi-source data acquisition unit is used to collect multi-modal monitoring data of the operating environment of the target optoelectronic device, and comprehensively obtain key parameter information reflecting the working state of the optoelectronic device. Then, a data preprocessing unit is used to perform preprocessing operations such as denoising and feature extraction on the collected multi-source monitoring data to generate a monitoring data set for device health status assessment. Next, the health status assessment unit 13 identifies the health degradation characteristics of the target optoelectronic device based on the monitoring data set, quantitatively assesses the health status of the target optoelectronic device, and promptly detects signs of faults. Afterwards, the fault diagnosis unit analyzes the health status assessment results, identifies device faults, and generates a fault diagnosis report. Compared with the prior art,
[0013] This application adopts an end-to-end data-driven approach to fully mine and integrate multi-source heterogeneous monitoring data of the operating conditions of optoelectronic devices, and constructs a full-process fault detection framework from data collection, feature extraction, health assessment to fault diagnosis, thereby improving the comprehensiveness, accuracy and efficiency of optoelectronic device fault detection.
[0014] After introducing the basic principles of the present application, the non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the drawings in the specification.
[0015] like Figure 1 As shown, the embodiment of the present application provides an optoelectronic device fault detection platform based on multi-source data fusion, which includes:
[0016] The multi-source data acquisition unit 11 is used to acquire multi-source monitoring data of the operating environment of the target optoelectronic device, wherein the multi-source monitoring data includes a plurality of parameter monitoring results reflecting the operating state of the target optoelectronic device.
[0017] Specifically, the multi-source monitoring data includes multiple parameter monitoring results reflecting the working state of the target optoelectronic device, including but not limited to environmental and working parameters such as temperature, humidity, vibration, voltage, current, and light intensity. Through the multi-source monitoring data, the actual working state and environmental conditions of the target optoelectronic device are fully reflected. The multi-source data acquisition unit 11 includes a variety of sensors and data acquisition devices, such as temperature sensors, humidity sensors, acceleration sensors, voltmeters, ammeters, light intensity sensors, etc. These devices continuously collect relevant parameters at a predetermined sampling frequency and transmit the collected data to the subsequent data preprocessing unit 12.
[0018] By collecting multi-source monitoring data, comprehensive working status information of the target optoelectronic device can be obtained, providing basic data support for subsequent data processing and fault diagnosis. Compared with a single data source, the multi-source data collection method can more accurately reflect the actual working condition of the device and help improve the accuracy and reliability of fault detection.
[0019] The data preprocessing unit 12 is used to preprocess the multi-source monitoring data to generate a monitoring data set for device health status assessment.
[0020] Specifically, the multi-source monitoring data acquired by the multi-source data acquisition unit 11 has a wide range of sources and various data formats, such as analog data, digital data, text data, image data, etc., and these data may have noise interference, data missing, data redundancy and other problems. In order to ensure the reliability and consistency of the monitoring data, the original collected multi-source monitoring data is preprocessed as necessary before the health status assessment is performed. Among them, the data preprocessing unit 12 adopts a variety of data preprocessing technologies, including but not limited to data cleaning, data normalization, data filling, data compression, etc. Through data cleaning, noise interference and abnormal values in multi-source monitoring data are filtered out; through data normalization, monitoring data of different dimensions and numerical ranges are unified to the same scale; through data filling, missing data are estimated and supplemented; through data compression, data redundancy is reduced, and data storage and transmission overhead are reduced. After the preprocessing operation, the data preprocessing unit 12 converts the heterogeneous original monitoring data of multiple sources into a structured monitoring data set, thereby improving the efficiency and accuracy of fault detection.
[0021] The health status assessment unit 13 is used to perform health status assessment on the target optoelectronic device according to the monitoring data set to obtain a health status assessment result.
[0022] Specifically, first, the health status assessment unit 13 obtains the preprocessed monitoring data set from the data preprocessing unit 12 through the data interface, which contains a time series record reflecting the performance parameters of the target optoelectronic device under different working conditions, such as working current, light output power, spectral characteristics, etc. After receiving the monitoring data set, the health status assessment unit 13 performs further feature extraction and statistical analysis on it. For example, the mean, variance, change trend and other statistics of different performance parameters are calculated, the correlation between the parameters is analyzed, and a multi-dimensional device performance feature vector is constructed to comprehensively characterize the health status of the device. Then, the health status assessment unit 13 inputs the extracted device performance feature vector into the pre-trained device health status assessment model, starts the model inference calculation, and obtains the health status assessment result. The device health status assessment model is constructed based on machine learning algorithms, such as support vector machines, random forests, convolutional neural networks, etc.
[0023] The fault diagnosis unit 14 is used to determine whether the target optoelectronic device is in a fault state based on the health status assessment result, and generate a fault diagnosis result when the optoelectronic device is in a fault state.
[0024] Specifically, first, the fault diagnosis unit 14 receives the health status assessment result output by the health status assessment unit 13 in real time through the data interface, which includes the qualitative judgment and quantitative indicator description of the current health status of the target optoelectronic device, such as health level, key performance parameter prediction value, degradation degree, etc. After receiving the health status assessment result, the fault diagnosis unit 14 compares it with the pre-set device fault threshold for judgment. The device fault threshold is obtained by experts in the optoelectronic field based on factors such as material properties, structural parameters, process level, etc. of the target optoelectronic device, combined with historical operation data and failure case analysis, to characterize the boundary conditions for device performance degradation to an unacceptable degree or functional failure. When the health status assessment result does not reach the fault threshold, the fault diagnosis unit 14 considers that the device is still in normal working condition and there is no need to generate fault diagnosis information. However, when the health status assessment result exceeds the fault threshold, such as a key performance parameter of the target optoelectronic device seriously deviates from the normal range, or the comprehensive degradation index reaches a dangerous level, the fault diagnosis unit 14 will determine that the optoelectronic device is currently in a fault state.
[0025] Through online real-time monitoring of data and health status assessment, performance anomalies and functional degradation of target optoelectronic devices can be discovered in a timely manner, providing practical support for device maintenance decisions and fault prevention, and effectively improving the operational reliability of optoelectronic devices.
[0026] Furthermore, the embodiment of the present application also includes:
[0027] Construct a health status assessment model for target optoelectronic devices;
[0028] Processing the monitoring data set according to the input format of the health status assessment model to obtain a standard data set;
[0029] The standard data set is input into the health status assessment model, and a health status assessment result is output.
[0030] In a feasible implementation, first, a health status assessment model of the target optoelectronic device is constructed to achieve quantitative analysis of the device health status. In order to give full play to the value of multi-source monitoring data, the health status assessment model is constructed using data-driven methods such as machine learning, and integrates the knowledge of the physical mechanism of optoelectronic devices. The input is the monitoring parameters that reflect the performance and working conditions of the device, and the output is a comprehensive indicator that characterizes the health status of the device. Specifically, according to the structural principles, material properties, functional requirements, etc. of the target optoelectronic device, the monitoring parameters that have an important impact on the health status of the device are determined, and the sampling records of the relevant parameters are screened from the historical operation data to form a sample data set; then, the sample data set is analyzed and mined to extract statistical feature quantities that reflect the degradation trend of the device, such as mean, variance, correlation coefficient, etc.; then, a machine learning model, such as support vector machine, convolutional neural network, etc., is selected, and the model parameters are learned and tuned using training samples; then, the model performance is evaluated on the test set, and the model is verified and confirmed in combination with expert experience, so as to obtain a health status assessment model.
[0031] Then, since the monitoring data set does not meet the input requirements of the health status assessment model, after obtaining the monitoring data set, the monitoring data set is processed according to the input format of the health status assessment model to make it meet the input requirements of the health status assessment model and obtain the standard data set. After that, the standard data set is input into the health status assessment model to obtain the device health status diagnosis result. Through model reasoning calculation, the device health degradation information contained in the standard data set is quantified into an intuitive health status indicator to obtain the health status assessment result.
[0032] By constructing a health status assessment model and extracting device degradation characteristics from multi-source monitoring data, real-time quantitative assessment and trend prediction of device health status are achieved, potential fault hazards are discovered in a timely manner, data support is provided for condition-based repair and preventive maintenance, and the long-term trouble-free operation capability of optoelectronic devices is improved.
[0033] Further, such as Figure 2 As shown, the embodiment of the present application also includes:
[0034] Constructing a device physical mechanism sub-model, wherein the device physical mechanism sub-model establishes a quantitative correlation between device performance parameters and environmental stress factors based on the physical structure and material properties of the target optoelectronic device;
[0035] Constructing a device degradation prediction sub-model, wherein the device degradation prediction sub-model extracts statistical feature quantities reflecting the degradation trend based on the device degradation history data, and performs a probability prediction of the device degradation process;
[0036] The device physical mechanism sub-model and the device degradation prediction sub-model are integrated to obtain a health status assessment model.
[0037] In a preferred embodiment, the health status assessment model consists of two sub-models, namely the device physical mechanism sub-model and the device degradation prediction sub-model, which characterize the health degradation mechanism of the device from different angles and improve the accuracy and reliability of the health status assessment through model fusion.
[0038] The device physical mechanism sub-model is constructed based on the physical structure parameters and material characteristic parameters of the target optoelectronic device. Its purpose is to establish a quantitative mapping relationship between the performance parameters of the device and the stress factors of the environment in which the device is located. By analyzing the internal physical processes of the device, such as carrier transport and photogenerated carrier recombination, the influence of changes in environmental stress (such as temperature, humidity, voltage, etc.) on the device performance is quantitatively described. The device degradation prediction sub-model is constructed based on the historical degradation data of the device. By collecting monitoring data on the performance degradation of the device under actual operating conditions, statistical characteristics that can reflect the degradation trend, such as aging rate, threshold current attenuation, etc., are extracted from it. Then, the degradation data is modeled using a probabilistic model to quantitatively predict the degradation process of the device in the future and its uncertainty.
[0039] After constructing the device physical mechanism sub-model and the device degradation prediction sub-model, the device physical mechanism sub-model and the device degradation prediction sub-model are integrated by adopting uncertainty modeling methods such as Bayesian inference. The device physical mechanism sub-model can provide prior knowledge and physical constraints for degradation prediction, while the device degradation prediction sub-model can perform data-driven corrections to the device physical mechanism sub-model, so that the two complement each other and work together to improve the accuracy and reliability of health status assessment.
[0040] By constructing a health status assessment model from two dimensions: the internal physical mechanism of the device and the external degradation law, the interpretability and long-term stability of the assessment results are significantly improved. This makes full use of prior physical knowledge and massive monitoring data, overcomes the limitations of a single model, and provides support for fault detection of optoelectronic devices.
[0041] Furthermore, the embodiment of the present application also includes:
[0042] Acquiring physical structure parameters and material characteristic parameters of the target optoelectronic device;
[0043] Based on the physical structure parameters and the material characteristic parameters, establishing a multi-physical field numerical model inside the target optoelectronic device;
[0044] For the multi-physics field numerical model, multi-physics field coupling numerical simulation analysis is performed to obtain the evolution law of device performance parameters under different environmental stress conditions;
[0045] A quantitative mapping relationship between device performance parameters and environmental stress factors is established to obtain the device physical mechanism sub-model.
[0046] In a preferred embodiment, first, the physical structure parameters and material property parameters of the target optoelectronic device are obtained. The physical structure parameters include the geometric size, electrode layout, waveguide design, etc. of the target optoelectronic device, which determine the boundary conditions of the physical fields such as the electric field distribution and the light field distribution inside the target optoelectronic device. The material property parameters include the bandgap width, carrier mobility, refractive index, etc. of the active area material, which determine the rate equations of the physical processes such as carrier transport and photogenerated carrier recombination inside the target optoelectronic device. Then, based on the obtained physical structure parameters and material property parameters, a multi-physical field numerical model inside the device is established, including electric field, light field, thermal field, stress and strain field, etc. For different physical fields, the corresponding partial differential equations are selected for mathematical description, such as Maxwell equations, heat conduction equations, etc. Then, using numerical discretization methods such as finite elements and finite differences, the control equations are discretized in space and time on the geometric model of the target optoelectronic device to form a large-scale linear or nonlinear algebraic equation group, and a multi-physical field numerical model is established.
[0047] Afterwards, numerical simulation analysis of multi-physics coupling is carried out for the established multi-physics numerical model. Since the working process of optoelectronic devices involves the interaction and coupling of multiple physical fields, multi-physics coupling is used to perform joint numerical simulation. For example, by coupling the carrier transport equation and the optical waveguide transmission equation, the mutual influence of the carrier concentration distribution and the light intensity distribution is analyzed; by coupling the heat conduction equation and the stress-strain equation, the modulation effect of device heating and thermal stress distribution on the carrier mobility is analyzed. Through multi-physics coupling simulation, the evolution law of the physical process inside the device under different environmental temperature, humidity, voltage and other stress conditions is systematically analyzed to obtain the three-dimensional distribution and response characteristic curve of the device performance parameters. Then, based on the results of multi-physics coupling simulation, a quantitative mapping relationship between device performance parameters and environmental stress factors is established to form a device physical mechanism sub-model. Specifically, key physical quantities related to device performance degradation, such as the maximum temperature of the active area and the peak current density, are extracted from the results of the multi-physics field coupling simulation. Then, a mapping function from environmental stress parameters to key physical quantities is constructed through data-driven methods such as multivariate regression and machine learning. Through physical mechanism analysis, the key physical quantities are linked to device performance parameters such as the IV characteristics, spectral response, and frequency characteristics of the device, forming a quantitative mapping relationship between environmental stress and device performance, namely, the device physical mechanism sub-model.
[0048] By constructing a device physical mechanism sub-model that describes the intrinsic mechanism between environmental factors and device performance, it provides important theoretical tools and analytical means for fault detection of optoelectronic devices, significantly improving the accuracy of health status assessment.
[0049] Furthermore, the embodiment of the present application also includes:
[0050] Acquiring historical monitoring data of the target optoelectronic device, wherein the historical monitoring data includes device performance parameters and corresponding environmental stress parameters;
[0051] Extracting features from the historical monitoring data to obtain statistical feature quantities reflecting the degradation state of the device;
[0052] Based on the statistical characteristic quantity, constructing a Markov random process model of the device degradation state;
[0053] The Markov random process model is trained to establish a probability transfer matrix between the statistical feature quantity and the device degradation state, thereby obtaining a device degradation prediction sub-model.
[0054] In a preferred embodiment, first, the historical monitoring data of the target optoelectronic device is obtained, which comes from the accelerated life test of the device or the actual engineering application, including the device performance parameters of the device that evolve over time under different environmental stress conditions, such as light output power, threshold current, spectral width, etc., and the corresponding stress parameters such as ambient temperature, humidity, and driving current. To ensure data quality, the historical monitoring data is preprocessed, such as removing outliers, interpolating missing values, etc. At the same time, in order to enhance the generalization of the model, device monitoring data from different batches and different application scenarios are collected to cover a variety of degradation modes and stress conditions. Subsequently, the acquired historical monitoring data is feature extracted to obtain statistical feature quantities that can reflect the degradation state of the device, including the mean, variance, degradation rate, etc. of the device performance parameters, and the correlation coefficients between different performance parameters. The statistical feature quantities are extracted by mathematical and statistical analysis of the historical monitoring data, which can filter out the influence of random noise and condense the inherent laws of the device degradation process. In addition, signal processing methods such as wavelet analysis and independent component analysis can also be used to dig out the degradation pattern characteristics hidden in the time series data.
[0055] Afterwards, based on the extracted statistical features, a Markov random process model that can describe the evolution of the device degradation state is constructed. The device degradation process is modeled as a Markov random process, that is, the health state of the device is divided into a finite number of discrete levels, such as normal, mild degradation, severe degradation, failure, etc.; then it is assumed that the transition of the device between these states follows the Markov property, that is, the current degradation state of the device is only related to the state of the previous moment, and has nothing to do with the earlier historical state. In the actual modeling, according to the distribution law of the statistical feature quantity, the state division method and the state transition trigger condition are selected to accurately characterize the step-like evolution characteristics of device degradation. Afterwards, the constructed Markov degradation model is parameterized and trained using historical monitoring data, and a state transition probability matrix between the statistical feature quantity and the device degradation state is established to form a device degradation prediction sub-model. The state transition probability matrix describes the probability law of the device transitioning between different degradation states, and is the core parameter of the Markov degradation model. Specifically, we use parameter identification methods such as maximum likelihood estimation and Bayesian estimation to learn the prior distribution and posterior distribution of the state transition probability matrix from historical monitoring data; then we use iterative optimization methods such as expectation maximization algorithm and Gibbs sampling algorithm to solve the optimal point estimate of the state transition probability matrix. Through data-driven parameter learning, we can fully explore the device degradation laws contained in massive monitoring data, overcome the subjectivity and limitations of traditional empirical estimation, and significantly improve the objectivity and applicability of the degradation prediction sub-model.
[0056] By utilizing data analysis and machine learning techniques, degradation features are extracted from massive historical monitoring data of optoelectronic devices, and a degradation prediction model based on Markov random processes is constructed. This enables probabilistic prediction of device health status and remaining life estimation, providing key support for data-based predictive maintenance decisions.
[0057] Furthermore, the embodiment of the present application also includes:
[0058] Inputting the standard data set into the device physical mechanism sub-model to obtain a physical mechanism evaluation result of the current health state of the target optoelectronic device, wherein the physical mechanism evaluation result includes a device performance parameter prediction value and a physical degradation index;
[0059] Inputting the standard data set into the device degradation prediction sub-model to obtain a statistical evaluation result of the current health state of the target optoelectronic device, wherein the statistical evaluation result includes a device degradation state probability and a statistical degradation index;
[0060] The physical mechanism evaluation result and the statistical evaluation result are comprehensively analyzed to obtain quantitative indicators of the current health state of the target optoelectronic device and a probability prediction of the future degradation state of the device, thereby generating a health state evaluation result.
[0061] In a feasible implementation, first, the standard data set is input into the device physical mechanism sub-model to obtain the health status assessment result based on the physical mechanism analysis. Specifically, the physical mechanism sub-model uses the quantitative mapping relationship between the embedded device performance parameters and the environmental stress factors based on the input environmental stress parameters that reflect the current working state of the device, such as temperature, humidity, driving current, etc., to obtain the theoretical prediction values of the device performance parameters under the current state, such as light output power, quantum efficiency, spectral line width, etc. At the same time, the device physical mechanism sub-model diagnoses and analyzes the impact of the current environmental stress conditions on the physical processes inside the device, and quantitatively gives the key indicators such as the device active area temperature, carrier concentration, defect state density, etc. that reflect the degree of physical degradation of the device. The device performance parameter prediction values and physical degradation indicators together constitute the physical mechanism assessment results of the current health state of the device.
[0062] At the same time, the standard data set is input into the device degradation prediction sub-model to obtain the health status assessment results based on the statistical analysis of the degradation data. The degradation prediction sub-model extracts the key statistical feature quantities that reflect the degradation state of the device in the standard data set, such as the degradation rate, variance, and correlation coefficient of performance parameters. Then, using the pre-trained Markov degradation model, the most likely degradation state of the device is obtained based on the matching probability between the current statistical feature quantity and each degradation state, and the probability of staying in the degradation state is given. At the same time, the device degradation prediction sub-model calculates the statistical degradation indicators of the device, such as the expected value of the degradation state, the mean and variance of the degradation process, etc., to quantitatively evaluate the overall degradation level of the device. The device degradation state probability and the statistical degradation indicators together constitute the statistical evaluation results of the current health state of the device.
[0063] Afterwards, the physical mechanism evaluation results and the degradation statistical evaluation results of the device are fused to generate quantitative analysis indicators of the current health state of the device and probability prediction of the future degradation state, forming the final health state evaluation report. Among them, uncertainty analysis methods such as Bayesian inference are used to quantify and fusion-propagate the uncertainties of the two types of evaluation results. For example, the predicted value of the performance parameter in the physical mechanism evaluation can be used as the prior information for the prediction of the degradation state, and the Bayesian update is performed with the posterior probability of the degradation state obtained based on statistical data to obtain the fused posterior probability distribution of the device degradation state; the physical degradation index in the physical mechanism evaluation can be weighted averaged with the statistical degradation index to obtain a comprehensive quantitative index of the current degradation level of the device. After the deep fusion of the two types of evaluation results is completed, a health state evaluation report containing the device health state level, degradation degree index, remaining life distribution, etc. can be output, which intuitively presents the current health level of the device and the future degradation trend.
[0064] Furthermore, the embodiment of the present application also includes:
[0065] Comparing the health status assessment result with a preset device health status threshold, wherein the device health status threshold includes a device performance degradation threshold and a device degradation probability threshold;
[0066] When the predicted value of the device performance parameter exceeds the device performance degradation threshold, or the current device degradation state probability exceeds the device degradation probability threshold, it is determined that the target optoelectronic device is in a fault state, and a fault diagnosis result is generated.
[0067] In a feasible implementation, first, the device health status assessment result output by the health status assessment unit 13 is compared with the pre-set device health status threshold to achieve quantitative judgment of the device failure state. Among them, the device health status threshold includes two categories: one is the device performance degradation threshold, which is set for the key performance parameters of the device, such as optical output power, response time, etc. When the predicted value of the device performance parameter exceeds the corresponding degradation threshold, it can be determined that the device performance is seriously degraded and it is difficult to meet the normal working requirements; the other is the device degradation probability threshold, which is set for the probability of the degradation state of the device, such as dividing the device degradation state into four levels of normal, mild degradation, severe degradation, and failure. When the probability of the device being in the latter two degradation states exceeds the corresponding threshold, it can be determined that the device is severely degraded and is likely to fail in the short term. The setting of the device health status threshold needs to comprehensively consider the design indicators, process level, application scenarios and other factors of the device, and balance reliability and maintenance cost under the premise of taking into account the device performance margin and failure risk, and is given by experts in the field of optoelectronic devices.
[0068] According to the comparison between the device health status assessment result and the preset device health status threshold, it is determined whether the device is currently in a fault state and a fault diagnosis result is generated. Specifically, when the predicted value of the device performance parameter exceeds the corresponding device performance degradation threshold, or the device degradation state probability exceeds the corresponding device degradation probability threshold, it can be diagnosed that the target optoelectronic device is already in a fault state. At this time, the fault diagnosis unit 14 outputs fault alarm information, indicating the identification information such as the model, number, and location of the faulty device, and at the same time gives the fault status information such as the current performance parameters and degradation state of the faulty device to obtain the fault diagnosis result.
[0069] By monitoring the health status of devices online in real time, signs of failure can be discovered in advance before device performance seriously degrades and fails, so that faulty devices can be actively isolated and repaired, minimizing their impact on the system and improving the optoelectronic system's ability to operate without failure for a long period of time.
[0070] Furthermore, the embodiment of the present application further includes a fault warning unit, and the execution steps of the fault warning unit include:
[0071] receiving in real time the fault diagnosis result generated by the fault diagnosis unit;
[0072] Determine the fault level corresponding to the fault diagnosis result according to a preset fault classification rule;
[0073] According to the fault level, a corresponding fault warning strategy is adopted, a fault warning is performed based on the fault warning strategy, and a fault warning log is generated to record the time, level, sending object and processing progress of the fault warning to form a fault warning file.
[0074] In a preferred embodiment, a fault warning unit is introduced to perform graded warning and information recording according to the fault diagnosis result after detecting the fault of the target optoelectronic device.
[0075] First, the fault warning unit maintains a real-time communication connection with the fault diagnosis unit 14. When the fault diagnosis unit 14 detects that the target optoelectronic device has a fault and generates a fault diagnosis result, the fault warning unit immediately obtains relevant information. The fault diagnosis result includes the identification information of the faulty device (such as device model, number, installation location, etc.), fault state information (such as performance parameter over-limit range, device degradation state probability, etc.), and fault mode information (such as device open circuit, short circuit, performance index degradation, etc.). After receiving the fault diagnosis result, the fault warning unit extracts key information to provide a decision basis for the next step of fault classification and warning strategy formulation. Then, the fault warning unit performs a graded evaluation on the received fault diagnosis result according to the preset fault classification rules to determine the severity level of the fault. The fault classification rules are pre-formulated based on factors such as the functional importance of the optoelectronic device, the scope of the fault impact, and the consequences of the fault risk. The device fault is divided into multiple levels, such as general faults, major faults, severe faults, etc., and each level corresponds to different warning strategies and disposal requirements. For example, for general faults, only regular tracking and maintenance are required; for serious faults, immediate shutdown and isolation are required, and relevant responsible persons are notified to start the emergency plan process. The fault warning unit can automatically determine the severity level of the current fault by matching the key information in the fault diagnosis results with the fault classification rules, providing a basis for the selection of the warning strategy.
[0076] Afterwards, according to the determined fault level, the fault alarm unit automatically selects the corresponding fault alarm strategy, releases fault warning information to relevant personnel, and generates a fault alarm log to record the alarm-related information in detail. The fault alarm strategy is an information release plan formulated in advance for different fault levels, which clarifies the content, method, object and timeliness requirements of the alarm. For example, for general faults, the alarm content includes the location and identification information of the faulty device. The alarm method can be email or text message, and the recipient is the equipment administrator. The alarm frequency can be once a day; for serious faults, the alarm content must also include a specific description of the fault status and emergency response requirements. The alarm method must use strong reminder methods such as pop-up windows and voice broadcasts. The recipients cover on-site operators, equipment supervisors, and quality and safety managers. The alarm frequency must continue until the fault is resolved. At the same time, the fault alarm unit automatically generates and sends fault warning information according to the alarm strategy, and saves the alarm record in the form of a log, including the alarm time, alarm level, alarm content, alarm object, alarm method, fault handling status, etc., to form a complete fault alarm file to provide data support for subsequent fault response effect evaluation and optimization.
[0077] In summary, the optoelectronic device fault detection platform based on multi-source data fusion provided by the embodiments of the present application has the following technical effects:
[0078] The multi-source data acquisition unit is used to collect multi-source monitoring data of the operating environment of the target optoelectronic device. The multi-source monitoring data includes multiple parameter monitoring results reflecting the working status of the target optoelectronic device, thereby comprehensively describing the health status of the optoelectronic device and providing rich data support for subsequent fault detection. The data preprocessing unit is used to preprocess the multi-source monitoring data and generate a monitoring data set for device health status assessment, which is convenient for subsequent health status assessment and fault diagnosis. The health status assessment unit is used to assess the health status of the target optoelectronic device based on the monitoring data set, obtain the health status assessment results, timely discover the performance degradation trend of the device, accurately warn of potential fault risks, and provide a basis for predictive maintenance decisions. The fault diagnosis unit is used to determine whether the target optoelectronic device is in a fault state based on the health status assessment results. When the optoelectronic device is in a fault state, a fault diagnosis result is generated to guide fault location and maintenance, improve fault handling efficiency, reduce the impact time of the fault on the system, and improve system availability.
[0079] Any of the steps described above may be stored as computer instructions or programs in an unlimited computer memory, and may be called and recognized by an unlimited computer processor to implement any platform in the embodiments of the present application, without any unnecessary restrictions.
[0080] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.
Claims
1. Optoelectronic device fault detection platform based on multi-source data fusion, characterized by: include: A multi-source data acquisition unit, the multi-source data acquisition unit is used to collect multi-source monitoring data of the operating environment of the target optoelectronic device, the multi-source monitoring data includes a plurality of parameter monitoring results reflecting the working state of the target optoelectronic device; A data preprocessing unit, the data preprocessing unit is used to preprocess the multi-source monitoring data to generate a monitoring data set for device health status assessment; A health status assessment unit, the health status assessment unit being used to perform a health status assessment on the target optoelectronic device according to the monitoring data set to obtain a health status assessment result; A fault diagnosis unit is used to determine whether the target optoelectronic device is in a fault state based on the health status assessment result, and generate a fault diagnosis result when the optoelectronic device is in a fault state.
2. The optoelectronic device fault detection platform based on multi-source data fusion according to claim 1 is characterized in that: Performing a health status assessment on the target optoelectronic device according to the monitoring data set to obtain a health status assessment result includes: Construct a health status assessment model for target optoelectronic devices; Processing the monitoring data set according to the input format of the health status assessment model to obtain a standard data set; The standard data set is input into the health status assessment model, and a health status assessment result is output.
3. The optoelectronic device fault detection platform based on multi-source data fusion according to claim 2 is characterized in that: Construct a health status assessment model for target optoelectronic devices, including: Constructing a device physical mechanism sub-model, wherein the device physical mechanism sub-model establishes a quantitative correlation between device performance parameters and environmental stress factors based on the physical structure and material properties of the target optoelectronic device; Constructing a device degradation prediction sub-model, wherein the device degradation prediction sub-model extracts statistical feature quantities reflecting the degradation trend based on the device degradation history data, and performs a probability prediction of the device degradation process; The device physical mechanism sub-model and the device degradation prediction sub-model are integrated to obtain a health status assessment model.
4. The optoelectronic device fault detection platform based on multi-source data fusion according to claim 3 is characterized in that: Construct device physical mechanism sub-model, including: Acquiring physical structure parameters and material characteristic parameters of the target optoelectronic device; Based on the physical structure parameters and the material characteristic parameters, establishing a multi-physical field numerical model inside the target optoelectronic device; For the multi-physics field numerical model, multi-physics field coupling numerical simulation analysis is performed to obtain the evolution law of device performance parameters under different environmental stress conditions; A quantitative mapping relationship between device performance parameters and environmental stress factors is established to obtain the device physical mechanism sub-model.
5. The optoelectronic device fault detection platform based on multi-source data fusion according to claim 4 is characterized in that: Construct device degradation prediction sub-model, including: Acquiring historical monitoring data of the target optoelectronic device, wherein the historical monitoring data includes device performance parameters and corresponding environmental stress parameters; Extracting features from the historical monitoring data to obtain statistical feature quantities reflecting the degradation state of the device; Based on the statistical characteristic quantity, constructing a Markov random process model of the device degradation state; The Markov random process model is trained to establish a probability transfer matrix between the statistical feature quantity and the device degradation state, thereby obtaining a device degradation prediction sub-model.
6. The optoelectronic device fault detection platform based on multi-source data fusion according to claim 2 is characterized in that: Inputting the standard data set into the health status assessment model and outputting the health status assessment result includes: Inputting the standard data set into the device physical mechanism sub-model to obtain a physical mechanism evaluation result of the current health state of the target optoelectronic device, wherein the physical mechanism evaluation result includes a device performance parameter prediction value and a physical degradation index; Inputting the standard data set into the device degradation prediction sub-model to obtain a statistical evaluation result of the current health state of the target optoelectronic device, wherein the statistical evaluation result includes a device degradation state probability and a statistical degradation index; The physical mechanism evaluation result and the statistical evaluation result are comprehensively analyzed to obtain quantitative indicators of the current health state of the target optoelectronic device and a probability prediction of the future degradation state of the device, thereby generating a health state evaluation result.
7. The optoelectronic device fault detection platform based on multi-source data fusion according to claim 6 is characterized in that: Determining whether the target optoelectronic device is in a fault state based on the health status assessment result includes: Comparing the health status assessment result with a preset device health status threshold, wherein the device health status threshold includes a device performance degradation threshold and a device degradation probability threshold; When the predicted value of the device performance parameter exceeds the device performance degradation threshold, or the current device degradation state probability exceeds the device degradation probability threshold, it is determined that the target optoelectronic device is in a fault state, and a fault diagnosis result is generated.
8. The optoelectronic device fault detection platform based on multi-source data fusion according to claim 1, characterized in that: It also includes a fault warning unit, and the execution steps of the fault warning unit include: receiving in real time the fault diagnosis result generated by the fault diagnosis unit; Determine the fault level corresponding to the fault diagnosis result according to a preset fault classification rule; According to the fault level, a corresponding fault warning strategy is adopted, a fault warning is performed based on the fault warning strategy, and a fault warning log is generated to record the time, level, sending object and processing progress of the fault warning to form a fault warning file.
Citation Information
Patent Citations
Distribution transformer online monitoring method and system based on multi-source information fusion
CN111289829A
Equipment residual life prediction method and system based on Bayesian multi-source data fusion
CN114936504A
Transformer fault assessment method and system based on multi-source data graph fusion model
CN117332330A
Filter state monitoring and fault diagnosis method based on digital twin technology
CN117554265A
Screw air compressor fault prediction method based on multi-source data fusion
CN118378195A
Cited By
Generator performance detection method of multi-source heterogeneous big data platform, medium and equipment
CN119738716A
Generator performance detection method, medium and equipment of multi-source heterogeneous big data platform
CN119738716B
Method and device for detecting stability of optical coupling device
CN120253171A
Multi-parameter fusion non-all-plastic hook type commutator fault diagnosis method and system
CN120333530A
Electrolysis system fault intelligent detection method and system based on explosion-proof robot dog
CN120539619A
