Life prediction method, device and system for chain transmission system with multi-component coupling

Abstract

A life prediction method, device and system for a chain drive system with multiple components coupled, the method comprising: topologically decomposing multiple components of the chain drive system, extracting the fault relationship between the components of the system by combining historical fault data and expert knowledge; preprocessing the fault data to obtain a structured fault feature matrix; establishing a fault propagation network of the chain drive system by using a cloud model symbolic directed graph, and constructing a modular fault propagation path reasoning model based on the functional relationship and working condition changes of the system components; dynamically simulating and analyzing the fault propagation of the system in real time by searching for compatible paths and qualitative and quantitative conversion of cloud models; uniformly quantifying random uncertainty and cognitive uncertainty, and using a regression model to model the degradation trend of the chain drive system; optimizing model parameters, combining the first arrival time method to predict the remaining life of the chain drive system, and providing the confidence of the life prediction through confidence interval analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of life prediction of chain transmission systems, in particular to a life prediction method, device and system for multi-component coupled chain transmission systems. Specifically, it relates to a chain transmission system life prediction technology based on fault propagation model, multi-source data fusion and hybrid uncertainty quantification, which is particularly suitable for reliability analysis and life prediction in complex multi-component coupled chain transmission systems. Background Art

[0002] As a common mechanical transmission device, chain drive system is widely used in various types of automation equipment and intelligent equipment. The various components of the chain drive system operate for a long time in a complex working environment and face the risk of degradation, failure and performance decline. Since the chain drive system is usually composed of multiple components that cooperate with each other and there are complex fault propagation paths between the components, the traditional life prediction method based on a single component can no longer meet the needs of life prediction under complex working conditions.

[0003] Patent CN107967531A corrects the consumption process of the life cycle of components based on basic parameters such as the operating time of key components of scraper conveyor equipment, while taking into account external influencing factors such as different loads or different geological conditions and the characteristics of the components themselves, and finally obtains the equivalent time of component life consumption, and outputs the life and provides user-oriented maintenance, maintenance and replacement processing strategies to improve the safety and reliability of equipment operation. Patent CN112623631A uses a scraper equipped with a magnetic block in combination with a Hall sensor to detect chain breakage of a scraper conveyor, which can predict whether there is a hidden danger of chain breakage and whether the scraper has a swaying phenomenon, solving the problem of lack of real-time life prediction and chain break monitoring of scraper conveyors. However, these life prediction methods ignore the influence of multi-component coupling and cannot effectively handle the uncertainty of data, resulting in low prediction accuracy and inability to accurately predict the remaining life of the entire chain transmission system. Summary of the invention

[0004] The present invention proposes a life prediction method, device and system for a chain drive system with multiple components coupled, aiming to solve the shortcomings of the traditional chain drive system life prediction method in the face of complex component coupling, fault propagation and uncertainty factors. By integrating topological structure decomposition, data mining, fault propagation modeling, degradation trend analysis and remaining life prediction technologies, the present invention provides a comprehensive and efficient prediction model that can accurately predict the remaining life of the chain drive system.

[0005] To achieve the above object, the present invention provides a life prediction method for a chain transmission system with multiple components coupled, which specifically comprises the following steps:

[0006] S1: Mining the topological structure decomposition and fault relationship of the chain transmission system; performing topological structure decomposition on multiple components of the chain transmission system, combining historical fault data and expert knowledge, and using association rule mining algorithms to extract the fault relationships between the components of the system.

[0007] S2: Preprocess the fault data to generate a structured fault feature matrix; Preprocess the fault data, including data cleaning, noise processing and data reduction, to obtain a structured fault feature matrix to provide data support for subsequent fault propagation modeling.

[0008] S3: Model the fault propagation network of the chain drive system and construct a modular fault propagation path reasoning model; use the cloud model symbolic directed graph to establish the fault propagation network of the chain drive system, and construct a modular fault propagation path reasoning model based on the functional relationship of system components and operating condition changes.

[0009] S4: Dynamic simulation and real-time analysis of fault propagation; through compatible path search and qualitative and quantitative conversion of cloud models, dynamic simulation and real-time analysis of system fault propagation are performed.

[0010] S5: Model the degradation trend and uniformly quantify the random and epistemic uncertainties; Uniformly quantify the random uncertainty and epistemic uncertainty, and use the regression model to model the degradation trend of the chain drive system.

[0011] S6: Predict the remaining life of the chain drive system, and give the confidence of the life prediction by combining the confidence interval analysis; optimize the model parameters through the parameter estimation algorithm, and predict the remaining life of the chain drive system by combining the first arrival time method, and give the confidence of the life prediction by confidence interval analysis.

[0012] Step S1 specifically includes:

[0013] The topological structure of multiple components of the chain transmission system is decomposed to construct a directed graph model of the system. Assume that the system consists of n components, each component Connected with other components through transmission relationship, component assembly , edge set Represents the fault propagation relationship between components. Through topological analysis, the functional dependencies and transmission structure between components can be identified.

[0014] Combined with historical fault data , for each fault record Processing, recording in time The fault state of the component when Representation Components Whether a fault has occurred. By using the association rule mining algorithm, the potential fault propagation relationship between components is analyzed and rules are generated. For each rule, calculate its support and confidence to identify the corresponding relationship between component failure and other component failures. Support represents the rule The frequency of occurrence, the confidence level is expressed in In case of occurrence Probability of occurrence.

[0015] Step S2 specifically includes:

[0016] Clean the fault data and process missing values. The fault data matrix is , each line Corresponding time point For missing fault data , column mean filling method is used to fill missing values ​​to ensure data completeness.

[0017] Each data point is smoothed by the moving average method of the sliding window to eliminate measurement errors or short-term fluctuations, making the data trend more stable and convenient for subsequent analysis. The fault data matrix is ​​reduced to A low-dimensional matrix, where , thereby reducing redundant information and retaining key features.

[0018] Step S3 specifically includes:

[0019] Construct a symbolic directed graph of the chain transmission system to represent the fault propagation relationship between the components in the system. Suppose the symbolic directed graph is ,in is a collection of component nodes, is the set of propagation relationship edges between components, is a set of fault propagation rules. Each edge Representation Components Failure of the component may result in of faults, and each edge has a corresponding propagation probability .

[0020] A fault propagation model of the chain drive system is constructed, and the membership function of the cloud model is used to describe the probability of component failure. The membership function is used to represent the fuzziness and uncertainty of the fault, where the expected fault value and standard deviation represent the expected level of component failure and the dispersion of the failure probability, respectively.

[0021] The fault propagation path is derived by using the inference mechanism. To target component The propagation path is , whose propagation probability is the product of the propagation probabilities of all edges in the path. The fault propagation probability of each path is calculated through the reasoning process, thereby identifying potential fault propagation paths in the system.

[0022] Step S4 specifically includes:

[0023] Through the compatible path search algorithm, starting from the fault source component, all possible fault propagation paths in the system are found. The propagation probability of each path is calculated by multiplying the propagation probabilities between all components on the path, which is used to identify each possible fault propagation path in the system.

[0024] The membership function of the cloud model is used to describe the failure probability of each component. By converting the membership function into the probability of failure, the fuzziness and uncertainty of component failure under different conditions are expressed.

[0025] Step S5 specifically includes:

[0026] The two types of uncertainty in the chain drive system are quantified uniformly using evidence theory. Specifically, random uncertainty and epistemic uncertainty are combined through evidence synthesis operation to obtain the total uncertainty.

[0027] The degradation trend of the chain drive system is modeled using the regression method to obtain the degradation model. Assume that the degradation state of the system at time t is , the input feature at the current moment is The regression model performs kernel mapping on the input features by introducing relevant vectors and obtains the regression equation:

[0028] (1)

[0029] in, is the regression coefficient, is the kernel mapping of the input features, is the noise term. In order to adapt to the dynamic characteristics of the system degradation process, the regression coefficient is adjusted by recursive updating, so that the degradation model can gradually adapt to the degradation trend of the system and accurately predict the future degradation state.

[0030] In order to better describe the randomness and uncertainty in the system degradation process, the additive noise model is introduced into the degradation model. It has a mean of zero and a variance of Gaussian distribution This noise model is used to simulate the random fluctuations in the system degradation process, and then quantify the system degradation process and predict possible future failure trends.

[0031] Step S6 specifically includes:

[0032] The degradation model parameters are optimized by parameter estimation algorithm; specifically, assuming that the degradation process of the chain drive system is composed of the hidden state Description, and the Cain state is related to the observed state of the system First, estimate the hidden state based on the current parameters Then, the expected likelihood function is maximized and the model parameters are updated. After multiple iterations, the optimal parameters can be converged to improve the prediction accuracy of the degradation model.

[0033] The first arrival time method is used to evaluate the system life, which describes the time required for the system to reach the final state from the initial state. In the life prediction of the chain drive system, it is assumed that is the remaining life, i.e., the time from the current state to the failure state. Assume is the time point when the system reaches the fault state, and the fault state is defined as the observed value Failure threshold exceeded Therefore, the remaining life It can be expressed as:

[0034] (2)

[0035] In order to quantify the reliability of the remaining life prediction, it is necessary to calculate the remaining life Confidence interval of ,in is the width of the confidence interval. The Monte Carlo method is used to simulate and generate the probability distribution of the remaining life, and the confidence interval is calculated based on this. The confidence interval is calculated as follows:

[0036] (3)

[0037] in, is the significance level, usually , corresponding to a 95% confidence level.

[0038] The second aspect of the present invention relates to a life prediction device for a chain transmission system with multiple components coupled, specifically comprising:

[0039] Chain drive data acquisition module, used to collect real-time working status data of each component in the chain drive system, including chain tension, chain speed, gear speed, working temperature, load and other physical quantities;

[0040] The multi-component coupling feature extraction module is used to perform multi-dimensional feature fusion on the collected data of each component, extract key feature parameters reflecting the overall operating status of the system based on the interaction relationship between different components, and then establish a working status model of multi-component coupling;

[0041] The life prediction model module is used to combine the multi-component coupling characteristics and predict the remaining life of the chain drive system based on the machine learning algorithm to evaluate the health status of the chain drive system;

[0042] Prediction result optimization module, used to correct and optimize the life prediction results, combining historical failure data, environmental changes and other factors to improve prediction accuracy and reliability;

[0043] Abnormal warning module, when the life prediction value is lower than the set threshold, it automatically sends out a warning signal to prompt relevant personnel to repair or replace the chain drive components to prevent system failure;

[0044] Remote monitoring and data transmission module, used to transmit prediction results and warning information to the remote monitoring platform in real time through wireless communication technology, so that relevant technicians can remotely monitor the health status of the chain drive system;

[0045] The data storage and analysis module is used to store the equipment's historical data, life prediction results and fault records, providing a basis for subsequent technical analysis and improvement.

[0046] The third aspect of the present invention relates to a life prediction system for a chain transmission system with multiple components coupled, specifically comprising:

[0047] Chain drive system status data acquisition equipment is used to collect real-time operating status data of components such as chains, gears, and bearings. The equipment includes temperature sensors, vibration sensors, load sensors, speed sensors, etc., which can monitor the working status of each component in real time and transmit the data to data processing and analysis equipment wirelessly or wiredly.

[0048] The multi-component coupling status data processing and analysis equipment fuses the collected data, predicts the remaining life of each component based on the multi-component coupling model, and estimates the life of the entire chain drive system. This includes using machine learning algorithms to analyze and model the collected raw data, considering the interaction between components, and obtaining the failure mode and life prediction model of each component through historical data training, so as to evaluate the overall health status of the chain drive system.

[0049] The life prediction result display device is used to display the life prediction results of each component and the entire system, and provide a visual analysis interface. The device includes a display or a remote monitoring platform, which is used to display the remaining life of each component and the entire system in real time, and provide visual analysis through charts, trend lines, etc., to help operation and maintenance personnel to grasp the health status of the system in a timely manner.

[0050] The early warning and alarm equipment can automatically trigger an alarm when the predicted result of the system's remaining life is close to the set threshold, and remind the operation and maintenance personnel to inspect or maintain the system through sound and light alarms, text message notifications, email notifications, etc.

[0051] The terminal device is connected to the above-mentioned equipment, receives and displays the life prediction results, alarm information and system status data, and supports remote monitoring and data query.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. The present invention can more comprehensively predict the remaining life of the chain drive system through multi-component coupling modeling and fault propagation path analysis. Traditional methods usually ignore the interaction between components, while this technology can comprehensively consider the dynamic changes of each component in the system, thereby improving the accuracy and reliability of life prediction.

[0054] 2. The present invention adopts uncertainty quantification technology, which can effectively deal with various uncertainties encountered in actual operation, such as environmental changes and component aging. The system can maintain a high prediction stability and provide reliable remaining life estimation even in complex and changing working environments.

[0055] 3. The present invention can timely detect potential faults and issue early warnings by real-time monitoring of system status and analysis of key parameters. This allows intervention before the problem develops into a serious fault, avoiding production interruptions or expensive repair costs.

[0056] 4. The present invention combines data analysis and machine learning to provide intelligent decision support to help maintenance personnel develop the best maintenance plan. By predicting the health of equipment, the system can optimize maintenance timing and resource allocation, reducing unnecessary maintenance costs.

[0057] 5. The technology of the present invention can be widely applied to various chain drive systems and can be customized according to different requirements. Its modular design enables the technology to be flexibly adapted to equipment in different fields and provides the possibility for future system expansion and upgrade.

[0058] 6. The accurate remaining life prediction of the present invention can avoid excessive maintenance and equipment downtime, reduce maintenance costs. Through early warning and optimized maintenance strategy, it can significantly reduce the losses caused by sudden failures and improve the overall economic benefits of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a schematic diagram of the method of the present invention;

[0060] Figure 2 is a schematic diagram of data preprocessing in the method of the present invention;

[0061] Figure 3 is a schematic diagram of uncertainty quantification and degradation trend modeling in the method of the present invention;

[0062] Figure 4 is a schematic diagram of the apparatus of the present invention;

[0063] Figure 5 is a schematic diagram of the system of the present invention. DETAILED DESCRIPTION

[0064] In order to make the technical solution of the present invention clearer, the technical solution provided by the present invention will be described in detail in conjunction with specific embodiments, and the present invention will be further described in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0065] Embodiment 1

[0066] Reference Figure 1 This embodiment provides a life prediction method for a chain transmission system with multiple components coupled, which specifically includes the following steps:

[0067] S1: The topological structure of multiple components of the chain drive system is decomposed, and the fault relationship between the components of the system is extracted using the association rule mining algorithm combined with historical fault data and expert knowledge.

[0068] Specifically, S11: perform topological structure decomposition on the chain transmission system. In the chain transmission system, assume that the system consists of components, each component It is connected to other components through transmission relationships. Through topological decomposition, the chain transmission system can be represented as a directed graph .

[0069] in: Represents a collection of components of a system, is a set of edges, representing the fault propagation relationship between components.

[0070] S12: Extract fault relationships using association rule mining algorithms. Based on historical fault data , each fault record , records the time of the component The fault state when Representation Components Whether a fault occurs.

[0071] Furthermore, the association rule mining algorithm FP-Growth is used to extract the fault relationship between components. , calculate support and confidence:

[0072] Support: indicates the rule Frequency of occurrence:

[0073] (1-1)

[0074] Confidence: Indicates In case of occurrence, Probability of occurrence:

[0075] (1-2)

[0076] These rules help identify which component failures will cause failures in other components, thus providing a basis for fault propagation modeling.

[0077] S2: Preprocess the fault data, including data cleaning, noise processing and data reduction, to obtain a structured fault feature matrix to provide data support for subsequent fault propagation modeling.

[0078] refer to Figure 2 ,The specific data preprocessing steps are, S21: cleaning the data and processing missing values. Assume that the fault data matrix for The matrix, each row Corresponding time point For missing fault data , using the column mean filling method:

[0079] (2-1)

[0080] S22: Use smoothing to reduce noise in the data. Assume that the data point Contains noise , smoothed by moving average:

[0081] (2-2)

[0082] in: is the size of the sliding window, is the smoothed data value.

[0083] S23: Use principal component analysis (PCA) to reduce the dimensionality of the data. Assumptions yes PCA reduces the fault feature matrix to principal components :

[0084] (2-3)

[0085] in, yes The principal component matrix of as input for subsequent modeling.

[0086] S3: Use the cloud model symbolic directed graph to establish the fault propagation network of the chain drive system, and build a modular fault propagation path reasoning model based on the functional relationship and operating condition changes of the system components.

[0087] Specifically, S31: construct a symbolic directed graph of the chain transmission system. Use the symbolic directed graph to represent the fault propagation relationship of the chain transmission system. Suppose the symbolic directed graph is .

[0088] in: Represents a collection of component nodes. represents the set of propagation relationship edges between components, is a set of fault propagation rules, where each edge express Failure may cause The fault propagation probability is .

[0089] S32: Construct a fault propagation model for chain drive systems. Through the membership function of the cloud model To describe the components Probability of failure:

[0090] (3-1)

[0091] in, is the expected fault value, is the standard deviation, which indicates the ambiguity and uncertainty of the fault.

[0092] S33: Implement the fault propagation path reasoning of the chain transmission system. Use the reasoning mechanism to deduce the fault propagation path, assuming that To target component The propagation path is , whose propagation probability is the product of the propagation probabilities of all edges in the path:

[0093] (3-2)

[0094] Through this reasoning process, the fault propagation probability of each path can be calculated and further potential fault propagation paths in the system can be identified.

[0095] S4: Through compatible path search and qualitative and quantitative transformation of cloud models, dynamic simulation and real-time analysis of system fault propagation are performed.

[0096] Specifically, S41: compatible path search and propagation probability calculation. Given a fault source , find all possible fault propagation paths through the compatible path search algorithm. For each path , its propagation probability It is obtained by multiplying the propagation probabilities of all edges in the path:

[0097] (4-1)

[0098] This process is used to identify each possible fault propagation path in the system and its corresponding propagation probability.

[0099] S42: Qualitative and quantitative transformation of cloud models. In the cloud model, membership function is used To describe each component The failure probability of each component is converted into the failure probability:

[0100] (4-2)

[0101] in, It is the membership function of the cloud model, which represents the ambiguity and uncertainty of component failure under specific conditions.

[0102] S5: The random uncertainty and epistemic uncertainty are quantified uniformly, and the degradation trend of the chain drive system is modeled using a regression model.

[0103] refer to Figure 3 ,The specific steps of uncertainty quantification and degradation trend modeling are ,S51: Quantify the uncertainty uniformly. In the chain transmission system, uncertainty is ,mainly divided into two categories: random uncertainty ( ): Random changes caused by factors such as environment, operating condition fluctuations, and component aging. The second is cognitive uncertainty ( ): Uncertainty due to expert knowledge, model assumptions, etc.

[0104] Furthermore, the two uncertainties are unified through the theory of evidence. Specifically, the total uncertainty is obtained by combining random uncertainty and epistemic uncertainty. , which is calculated as:

[0105] (5-1)

[0106] Among them, the symbol Represents a synthesis operation of evidence, used to combine uncertainties from different sources.

[0107] S52: Modeling the degradation trend of chain drive systems using the recursive correlation vector regression (RVSR) model. is the degraded state of the system at time t, is the input feature at the current moment. The RVSR model introduces the relevant vector To model:

[0108] (1)

[0109] in: is the regression coefficient, is the kernel mapping of the input features, is the noise term.

[0110] Furthermore, in order to better adapt to the dynamic degradation process, the regression coefficients are adjusted by recursive updating. :

[0111] (5-3)

[0112] in, is the learning rate, is the dynamic error term. Through recursive updating, the model can gradually adapt to the degradation trend of the system, thereby more accurately predicting the future degradation state.

[0113] S53: Processing the randomness and uncertainty of the quantitative degradation process. In order to better characterize the randomness and uncertainty of the system degradation process, a random process term is added to the degradation model. The present invention uses an additive noise model to describe uncertainty:

[0114] (5-4)

[0115] in, represents the noise term, which has a mean of zero and a variance of In this way, the degradation of the system can be quantified and its future failure trend can be estimated.

[0116] S6: The model parameters are optimized through parameter estimation algorithm, and the remaining life of the chain drive system is predicted by combining the first arrival time method. The confidence of the life prediction is given through confidence interval analysis.

[0117] Specifically, S61: Use the expectation maximization (EM) algorithm to optimize the model parameters. Assume that the degradation process of the chain drive system consists of a hidden state Description, and the Cain state is related to the observed state of the system The model parameters are estimated by maximizing the likelihood function, the goal is:

[0118] (6-1)

[0119] Specifically, the steps of EM algorithm to optimize model parameters include: first, estimating the hidden state according to the current parameters The posterior distribution of ; Then, update the model parameters by maximizing the expected likelihood function :

[0120] (6-2)

[0121] After multiple iterations, the EM algorithm can converge to the optimal parameters, thereby improving the prediction accuracy of the degradation model.

[0122] S62: Use the first arrival time method to predict the remaining life of the chain transmission system. For the chain transmission system, assume that the remaining life of the system is , which is the time required for the system to degenerate from the current state to the fault state. Assume is the time point at which the system reaches a fault state, and the fault state is defined as the observed value Failure threshold exceeded The remaining life of the system is It can be expressed as:

[0123] (2)

[0124] in, It is the fault judgment threshold, which means that the system is considered to have failed when it reaches a certain degradation level.

[0125] S63: Provide confidence in the life prediction through confidence interval analysis. In order to quantify the reliability of the remaining life prediction, it is necessary to provide the remaining life Confidence interval of ,in is the width of the confidence interval. The calculation of the confidence interval is simulated using the Monte Carlo method. By performing a large number of random sampling on the model, the probability distribution of the remaining life is obtained. Based on this distribution, the confidence interval of the remaining life can be calculated.

[0126] The confidence interval calculation formula is:

[0127] (3)

[0128] in, is the significance level, usually , corresponding to a 95% confidence level.

[0129] Embodiment 2

[0130] Reference Figure 4This embodiment provides a life prediction device for a chain transmission system with multiple components coupled. The device is configured in a chain transmission system monitoring device, and can monitor the working status of each component of the chain transmission system in real time, predict the remaining life of the system, and issue a warning signal in time.

[0131] Reference Figure 4 The device includes: a chain drive data acquisition module, a multi-component coupling feature extraction module, a life prediction model module, a prediction result optimization module, an abnormality warning module, a remote monitoring and data transmission module, and a data storage and analysis module.

[0132] The chain drive data acquisition module is used to collect the working status data of each component in the chain drive system, including: chain tension sensor: used to measure the actual tension of the chain and reflect the load condition of the chain; chain speed sensor: used to monitor the movement speed of the chain and obtain the dynamic characteristics of the chain operation; gear speed sensor: used to obtain the speed information of the gears in the chain drive system in order to analyze the gear wear and load conditions; temperature sensor: used to detect the working temperature of each component in the chain drive system, especially the temperature of the chain and gears, as an important basis for evaluating its working environment; load sensor: used to detect the working load of the chain drive system, reflecting the stress distribution of the system and the load status of each component.

[0133] The multi-component coupling feature extraction module is used to perform multi-dimensional feature fusion and analysis based on the collected data. This module uses data fusion technology to couple the status information of each component and extract key features that reflect the health of the entire chain drive system. Through comprehensive analysis of chain tension, speed, temperature, load and other data, a mathematical model of the interaction between multiple components is established, thereby providing comprehensive data support for system life prediction.

[0134] The life prediction model module is used to predict the remaining life of the chain drive system based on the coupling characteristics of multiple components. This module uses machine learning and data mining technology, combined with the working status data of each component, to build a life prediction model for the chain drive system. The model predicts the remaining service life of the chain drive system by analyzing the historical operation data of the system, the wear pattern of each component and the load change. At the same time, the prediction model can adjust the prediction results according to real-time data, making the prediction results more accurate and reliable.

[0135] The prediction result optimization module is used to correct and optimize the life prediction results. This module dynamically corrects the prediction results based on historical fault data, environmental changes, operating abnormalities and other factors. By comparing the difference between the actual life and the predicted life, the prediction model is continuously optimized to improve the prediction accuracy and stability. The optimization module will also automatically adjust the weight parameters in the prediction model according to changes in environmental conditions and working conditions to ensure the accuracy of the prediction results under different working conditions.

[0136] The abnormal warning module is used to automatically issue a warning signal when the predicted value of the system life is lower than the set threshold. When the predicted value of the remaining life of the chain drive system is lower than the preset threshold, the abnormal warning module will automatically generate warning information. This information includes the predicted value, the probability of failure, and the possible location of the failed component. The warning information is sent to the remote monitoring platform via wireless communication. Relevant personnel can receive real-time warnings and take measures such as maintenance or replacement of components in time to prevent major system failures.

[0137] The remote monitoring and data transmission module is used to transmit the equipment status information, life prediction results and warning information to the remote monitoring platform in real time. This module transmits the equipment's real-time data and warning information to the cloud server or remote monitoring center through wireless communication technology, which is convenient for technicians to conduct real-time monitoring. The remote monitoring platform provides real-time display of equipment health status and historical trend analysis through big data analysis and visualization technology, helping managers make scientific decisions.

[0138] The data storage and analysis module is used to store the historical data, life prediction results and fault records of the equipment. This module can store the operation data, life prediction information, warning events and other data of each component for a long time, providing a basis for subsequent technical analysis and improvement. The stored data will be used to compare and analyze historical fault data to improve the accuracy and adaptability of the prediction model.

[0139] Embodiment 3

[0140] Reference Figure 5 This embodiment provides a life prediction system for a chain transmission system with multiple components coupled, including: a chain transmission system status data acquisition device, a multiple component coupling status data processing and analysis device, a life prediction result display device, an early warning and alarm device, and a terminal device.

[0141] The chain drive system status data acquisition equipment consists of multiple sensors, including temperature sensors, vibration sensors, speed sensors and load sensors installed on the chain, gears and bearings. All sensors are connected to the central control unit and can collect and transmit the operating data of each component of the chain drive system in real time, such as temperature changes, vibration frequency, load fluctuations, etc.

[0142] The multi-component coupling state data processing and analysis equipment uses machine learning algorithms to process the real-time data of each component. First, through data cleaning and preprocessing steps, the noise data is removed, and then the operating data of each component is input into the multi-component coupling model. The model predicts the remaining life of each component by analyzing the mutual influence between the components, and calculates the remaining life of the entire chain drive system. In addition, during the data processing process, the working environment factors of each component and the changes in workload are taken into account, and the model is adjusted using an adaptive algorithm so that the prediction results are continuously optimized as the real-time data changes.

[0143] The life prediction result display device displays the life prediction results of each component in real time through the LCD screen and remote cloud platform, and draws trend charts, life distribution charts, etc., to help operators intuitively understand the health status of the system. The device can also support historical data query, which is convenient for analyzing and comparing system health changes at different time points.

[0144] Early warning and alarm equipment: When the remaining life of a component or the entire system approaches the preset safety threshold, the system will automatically trigger an alarm. Alarm methods include real-time voice reminders, pop-up reminders, SMS, emails, and APP notifications to ensure that maintenance personnel can take timely measures.

[0145] Terminal devices, such as smartphones, computers or industrial tablets, can connect to various components of the system and data processing and analysis equipment to view the system's operating data and life prediction results in real time. By supporting multi-platform access, operation and maintenance personnel can view the system status on-site or remotely, schedule maintenance work, and perform system maintenance and upkeep in a timely manner.

[0146] The above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention is described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A life prediction method for a chain drive system with multiple components coupled, comprising the following steps: S1: Topological structure decomposition and fault relationship of mining chain transmission system; The topological structure of the multi-components of the chain drive system is decomposed, and the fault relationship between the components of the system is extracted using the association rule mining algorithm combined with historical fault data and expert knowledge; S2: Preprocess the fault data to generate a structured fault feature matrix; Preprocess the fault data, including data cleaning, noise processing and data reduction, to obtain a structured fault feature matrix to provide data support for subsequent fault propagation modeling; S3: Model the fault propagation network of the chain drive system and construct a modular fault propagation path reasoning model; The fault propagation network of the chain drive system is established using the cloud model symbolic directed graph, and a modular fault propagation path reasoning model is constructed based on the functional relationship and working condition changes of the system components; S4: Dynamic simulation and real-time analysis of fault propagation: Through compatible path search and qualitative and quantitative conversion of cloud models, dynamic simulation and real-time analysis of fault propagation of the system are performed; S5: Modeling degradation trends and quantifying stochastic and epistemic uncertainties uniformly; The random uncertainty and epistemic uncertainty are quantified uniformly, and the degradation trend of the chain drive system is modeled using a regression model; specifically, the following are included: The two types of uncertainty in the chain transmission system are quantified uniformly by using evidence theory; specifically, the random uncertainty and epistemic uncertainty are combined through evidence synthesis operation to obtain the total uncertainty; The degradation trend of the chain drive system is modeled using the regression method to obtain the degradation model. The degradation state of , the input feature at the current moment is ; The regression model performs kernel mapping on the input features by introducing relevant vectors and obtains the regression equation: ， in, is the regression coefficient, is the kernel mapping of the input features, is a noise term; in order to adapt to the dynamic characteristics of the system degradation process, the regression coefficient is adjusted by recursive updating, so that the degradation model can gradually adapt to the degradation trend of the system and accurately predict the future degradation state; In order to better describe the randomness and uncertainty in the system degradation process, the additive noise model is introduced into the degradation model; the noise term It has a mean of zero and a variance of Gaussian distribution ; This noise model is used to simulate the random fluctuations in the system degradation process, and then quantify the system degradation process and predict possible future failure trends; S6: Predict the remaining life of the chain drive system, and give the confidence of the life prediction by combining the confidence interval analysis; optimize the model parameters through the parameter estimation algorithm, and predict the remaining life of the chain drive system by combining the first arrival time method, and give the confidence of the life prediction by confidence interval analysis.

2. The life prediction method for a chain transmission system with multiple components coupled as claimed in claim 1, characterized in that: Step S1 specifically includes: The topological structure of multiple components of the chain transmission system is decomposed to construct a directed graph model of the system; suppose the system consists of components, each component Connected with other components through transmission relationship, component assembly , edge set Represent the fault propagation relationship between components; identify the functional dependencies and transmission structure between components through topological analysis; Combined with historical fault data , for each fault record Processing, recording in time The fault state of the component when Representation Components Whether a fault has occurred; by using the association rule mining algorithm, analyze the potential fault propagation relationship between components and generate rules ; For each rule, calculate its support and confidence to identify the corresponding relationship between component failure and other component failures; support represents the rule The frequency of occurrence, the confidence level is expressed in In case of occurrence Probability of occurrence.

3. The life prediction method for a chain transmission system with multiple components coupled as claimed in claim 1, characterized in that: Step S2 specifically includes: Clean the fault data and process missing values; the fault data matrix is , each line Corresponding time point Fault records; for missing fault data , use the column mean filling method to fill in missing values ​​to ensure data completeness; Each data point is smoothed by the moving average method of the sliding window to eliminate measurement errors or short-term fluctuations, making the data trend more stable and convenient for subsequent analysis; The fault data matrix is ​​reduced to A low-dimensional matrix, where , thereby reducing redundant information and retaining key features.

4. The life prediction method for a chain transmission system with multiple components coupled as claimed in claim 1, characterized in that: Step S3 specifically includes: Construct a symbolic directed graph of the chain transmission system to represent the fault propagation relationship between the components in the system; let the symbolic directed graph be ,in is a collection of component nodes, is the set of propagation relationship edges between components, is a set of fault propagation rules; each edge Representation Components Failure of the component may result in of faults, and each edge has a corresponding propagation probability ; A fault propagation model for the chain drive system is constructed, and the membership function of the cloud model is used to describe the probability of component failure. The fuzziness and uncertainty of the fault are represented by the membership function, where the expected fault value and standard deviation represent the expected level of component failure and the dispersion of the failure probability, respectively. The fault propagation path is derived by using the inference mechanism. To target component The propagation path is , whose propagation probability is the product of the propagation probabilities of all edges in the path; the fault propagation probability of each path is calculated through the reasoning process, thereby identifying the potential fault propagation paths in the system.

5. The life prediction method for a chain transmission system with multiple components coupled as claimed in claim 1, characterized in that: Step S4 specifically includes: Through the compatible path search algorithm, starting from the fault source component, all possible fault propagation paths in the system are found; the propagation probability of each path is calculated by multiplying the propagation probabilities between all components on the path, which is used to identify each possible fault propagation path in the system; The membership function of the cloud model is used to describe the failure probability of each component. By converting the membership function into the probability of failure, the fuzziness and uncertainty of component failure under different conditions are expressed.

6. The life prediction method for a chain transmission system with multiple components coupled as claimed in claim 1, characterized in that: Step S6 specifically includes: The degradation model parameters are optimized by parameter estimation algorithm; specifically, assuming that the degradation process of the chain drive system is composed of the hidden state Description, and the Cain state is related to the observed state of the system Related; First, estimate the hidden state based on the current parameters The posterior distribution of the regression model is obtained by maximizing the expected likelihood function and updating the model parameters. After multiple iterations, the optimal parameters can be converged to improve the prediction accuracy of the degradation model. The first arrival time method is used to evaluate the system life, which describes the time required for the system to reach the final state from the initial state. In the life prediction of the chain drive system, it is assumed that is the remaining life, i.e., the time from the current state to the failure state; assuming is the time point when the system reaches the fault state, and the fault state is defined as the observed value Failure threshold exceeded Therefore, the remaining life It can be expressed as: ， In order to quantify the reliability of the remaining life prediction, it is necessary to calculate the confidence interval of the remaining life ,in is the width of the confidence interval; the Monte Carlo method is used to simulate and generate the probability distribution of the remaining life, and the confidence interval is calculated based on this; the calculation formula of the confidence interval is: ， in, is the significance level, , corresponding to a 95% confidence level.

7. A life prediction device for a chain transmission system with multiple components coupled, characterized in that: include: Chain drive data acquisition module, used to collect the working status data of each component in the chain drive system in real time, including chain tension, chain speed, gear speed, working temperature, and physical quantities of load; The multi-component coupling feature extraction module is used to perform multi-dimensional feature fusion on the collected data of each component, extract key feature parameters reflecting the overall operating status of the system based on the interaction relationship between different components, and then establish a working status model of multi-component coupling; The life prediction model module is used to combine the multi-component coupling characteristics and predict the remaining life of the chain drive system based on the machine learning algorithm to evaluate the health status of the chain drive system; Prediction result optimization module, used to correct and optimize the life prediction results, combining historical failure data and environmental change factors to improve prediction accuracy and reliability; Abnormal warning module, when the life prediction value is lower than the set threshold, it automatically sends out a warning signal to prompt relevant personnel to repair or replace the chain drive components to prevent system failure; Remote monitoring and data transmission module, used to transmit prediction results and warning information to the remote monitoring platform in real time through wireless communication technology, so that relevant technicians can remotely monitor the health status of the chain drive system; The data storage and analysis module is used to store the equipment's historical data, life prediction results and fault records, providing a basis for subsequent technical analysis and improvement.

8. A life prediction system for a chain drive system with multiple components coupled, characterized in that: include: Chain drive system status data acquisition equipment, used to collect real-time operating status data of chain, gear, bearing components; the equipment includes temperature sensors, vibration sensors, load sensors, speed sensors, which can monitor the working status of each component in real time and transmit the data to the data processing and analysis equipment by wireless or wired means; Multi-component coupling status data processing and analysis equipment, which fuses the collected data, predicts the remaining life of each component based on the multi-component coupling model, and estimates the life of the entire chain drive system; including using machine learning algorithms to analyze and model the collected raw data, considering the interaction between the components, and obtaining the failure mode and life prediction model of each component through historical data training, so as to realize the evaluation of the overall health status of the chain drive system; Life prediction result display equipment is used to display the life prediction results of each component and the entire system, and provide a visual analysis interface; the equipment includes a display or a remote monitoring platform, which is used to display the remaining life of each component and the entire system in real time, and provide visual analysis in the form of charts and trend lines to help operation and maintenance personnel to grasp the health status of the system in a timely manner; Early warning and alarm equipment can automatically trigger an alarm when the system's remaining life prediction result is close to the set threshold, and remind the operation and maintenance personnel to inspect or maintain the system through sound and light alarms, SMS notifications, and email notifications; The terminal device is connected to the above-mentioned equipment, receives and displays the life prediction results, alarm information and system status data, and supports remote monitoring and data query.

Images