Filter element failure early warning method and device based on multi-dimensional parameter analysis

Abstract

The application provides a filter core failure early warning method and device based on multi-dimensional parameter analysis, relates to the technical field of failure early warning, and comprises the following steps: obtaining multi-dimensional data records and real-time multi-dimensional data; performing vibration cumulative damage analysis based on the input of a real-time multi-dimensional data into a nonlinear damage accumulation model; performing risk trend evaluation according to a high-frequency vibration influence state and failure early warning; performing failure prediction according to multi-stage failure early warning signals; performing simulation maintenance on a virtual model to obtain a virtual failure early warning node; judging the post-position effectiveness of the virtual failure early warning node on a failure prediction node, and determining whether to perform filter core maintenance according to a judgment result. Through the application, the technical problem that the filter core failure prediction accuracy is insufficient due to the inability to effectively integrate multi-dimensional parameters to comprehensively evaluate the filter core state in the prior art can be solved, the technical target of comprehensive evaluation of filter core multi-dimensional parameters is achieved, and the technical effect of accurately predicting the filter core failure time is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of failure warning, and particularly relates to a filter element failure warning method and device based on multi-dimensional parameter analysis. BACKGROUND

[0002] As a key filter component in industrial equipment, filter elements are widely used in fluid purification, air filtration and other fields. By filtering out impurities and particulate matter, the equipment is protected from pollutants, thereby prolonging the service life of the equipment.

[0003] At present, most of the existing technologies cannot effectively combine multi-dimensional parameters (such as temperature, pressure, vibration, etc.) for comprehensive evaluation of the health status of filter elements. The failure of filter elements is often the result of the combined action of multiple factors, and the monitoring of a single parameter cannot accurately capture the actual health status of the filter element. Therefore, the existing technologies have great limitations in the accuracy and real-time performance of filter element failure prediction, and cannot meet the needs of modern industry for efficient operation and maintenance management of equipment.

[0004] In summary, in the existing technology, there is a technical problem that the accuracy and real-time performance of filter element failure prediction are insufficient due to the inability to effectively integrate multi-dimensional parameters for comprehensive evaluation of the health status of filter elements, which further affects the efficient operation and maintenance management of equipment. SUMMARY

[0005] The purpose of the present application is to provide a filter element failure warning method and device based on multi-dimensional parameter analysis, to solve the technical problem in the existing technology that the accuracy and real-time performance of filter element failure prediction are insufficient due to the inability to effectively integrate multi-dimensional parameters for comprehensive evaluation of the health status of filter elements, which further affects the efficient operation and maintenance management of equipment.

[0006] In view of the above problems, the present application provides a filter element failure warning method and device based on multi-dimensional parameter analysis.

[0007] In a first aspect, the application provides a filter core failure warning method based on multi-dimensional parameter analysis, which is realized by a filter core failure warning device based on multi-dimensional parameter analysis, and includes the following steps: an interactive multi-modal sensor obtains multi-dimensional data records and real-time multi-dimensional data of a target filter core, the multi-dimensional data records include high-frequency micro-vibration records and corresponding multi-dimensional parameter records, and the multi-dimensional parameter records have a time duration identifier; a nonlinear damage accumulation model is constructed according to the high-frequency micro-vibration records, the multi-dimensional parameter records and the time duration identifier, vibration cumulative damage analysis is performed based on the real-time multi-dimensional data input into the nonlinear damage accumulation model, and a high-frequency vibration influence state of the target filter core is obtained; risk trend evaluation is performed according to the high-frequency vibration influence state, multi-level risk trends are obtained, failure warning is performed based on the multi-level risk trends, and multi-level failure warning signals are obtained; failure prediction is performed according to the multi-level failure warning signals, and a failure prediction node is obtained; a virtual model of the target filter core is constructed, simulation maintenance is performed on the virtual model, and a virtual failure warning node is obtained; and it is judged whether the virtual failure warning node is post-effective to the failure prediction node, and whether filter core maintenance is performed is determined according to a judgment result.

[0008] In a second aspect, the application also provides a filter core failure warning device based on multi-dimensional parameter analysis, which is used to execute the filter core failure warning method based on multi-dimensional parameter analysis as described in the first aspect, and includes the following modules: a multi-dimensional data record obtaining module, which is used for an interactive multi-modal sensor to obtain multi-dimensional data records and real-time multi-dimensional data of a target filter core, the multi-dimensional data records include high-frequency micro-vibration records and corresponding multi-dimensional parameter records, and the multi-dimensional parameter records have a time duration identifier; a high-frequency vibration influence state obtaining module, which is used for constructing a nonlinear damage accumulation model according to the high-frequency micro-vibration records, the multi-dimensional parameter records and the time duration identifier, performing vibration cumulative damage analysis based on the real-time multi-dimensional data input into the nonlinear damage accumulation model, and obtaining a high-frequency vibration influence state of the target filter core; a multi-level failure warning signal obtaining module, which is used for performing risk trend evaluation according to the high-frequency vibration influence state, obtaining multi-level risk trends, performing failure warning based on the multi-level risk trends, and obtaining multi-level failure warning signals; a failure prediction node obtaining module, which is used for performing failure prediction according to the multi-level failure warning signals, and obtaining a failure prediction node; a virtual failure warning node obtaining module, which is used for constructing a virtual model of the target filter core, performing simulation maintenance on the virtual model, and obtaining a virtual failure warning node; and a filter core maintenance module, which is used for judging whether the virtual failure warning node is post-effective to the failure prediction node, and determining whether filter core maintenance is performed according to a judgment result.

[0009] The one or more technical solutions provided in the application have at least the following technical effects or advantages:

[0010] The multi-dimensional data record and real-time multi-dimensional data of the target filter element are obtained through the interactive multi-modal sensor, the multi-dimensional data record includes a high-frequency micro-vibration record and a corresponding multi-dimensional parameter record, the multi-dimensional parameter record has a duration identifier; a non-linear damage accumulation model is constructed according to the high-frequency micro-vibration record, the multi-dimensional parameter record and the duration identifier, vibration cumulative damage analysis is performed based on the real-time multi-dimensional data input into the non-linear damage accumulation model, and a high-frequency vibration influence state of the target filter element is obtained; risk trend assessment is performed according to the high-frequency vibration influence state, a multi-level risk trend is obtained, and multi-level failure warning signals are obtained based on the multi-level failure warning signals; failure prediction is performed according to the multi-level failure warning signals, and a failure prediction node is obtained; a virtual model of the target filter element is constructed, and a virtual failure warning node is obtained by simulating and maintaining the virtual model; the post-effectiveness of the virtual failure warning node on the failure prediction node is judged, and it is determined whether to perform filter element maintenance according to the judgment result, thereby achieving the technical goal of comprehensive evaluation and real-time monitoring of multi-dimensional parameters of the filter element, and achieving the technical effect of accurately predicting the failure time of the filter element and improving the operation reliability and maintenance efficiency of the equipment.

[0011] The above description is only a summary of the technical solutions of the application. In order to more clearly understand the technical means of the application, the specific embodiments of the application can be implemented according to the content of the specification. In order to make the above and other purposes, characteristics and advantages of the application more obvious and easy to understand, the following specific embodiments of the application are described. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the application, nor is it intended to limit the scope of the application. Other features of the application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by those skilled in the art without creating any creative labor on the basis of the provided drawings.

[0013] Figure 1 The flowchart of the filter element failure warning method based on multi-dimensional parameter analysis of the application;

[0014] Figure 2 The structural schematic diagram of the filter element failure warning device based on multi-dimensional parameter analysis of the application.

[0015] Explanation of reference signs:

[0016] The multi-dimensional data record obtaining module 11, the high-frequency vibration influence state obtaining module 12, the multi-stage failure early warning signal obtaining module 13, the failure prediction node obtaining module 14, the virtual failure early warning node obtaining module 15, and the filter element maintenance module 16. DETAILED DESCRIPTION

[0017] The application provides a filter element failure early warning method and device based on multi-dimensional parameter analysis, which solves the technical problem in the prior art that the accuracy and real-time performance of filter element failure prediction are insufficient due to the inability to effectively integrate multi-dimensional parameters to comprehensively evaluate the health status of the filter element, and further affect the efficient operation and maintenance management of equipment. The technical target of comprehensive evaluation and real-time monitoring of multi-dimensional parameters of the filter element is achieved, and the technical effect of accurately predicting the failure time of the filter element and improving the operation reliability and maintenance efficiency of the equipment is achieved.

[0018] Hereinafter, the technical solutions in the application will be described clearly and completely with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments of the application. It should be understood that the application is not limited by the example embodiments described herein. Based on the embodiments of the application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the application. In addition, it should be noted that, for the convenience of description, only parts related to the application are shown in the drawings, not all.

[0019] Embodiment one, please refer to the attached Figure 1 The application provides a filter element failure early warning method based on multi-dimensional parameter analysis, which is applied to a filter element failure early warning device based on multi-dimensional parameter analysis, and specifically includes the following steps:

[0020] Step one: interactive multi-modal sensors obtain multi-dimensional data records and real-time multi-dimensional data of a target filter element, the multi-dimensional data records include high-frequency micro-vibration records and corresponding multi-dimensional parameter records, and the multi-dimensional parameter records have a time duration identifier.

[0021] Specifically, the interactive multi-modal sensor obtains multi-dimensional data records and real-time multi-dimensional data of the target filter during operation by simultaneously using multiple types of sensors such as acceleration sensors, acoustic sensors, optical fiber sensors, and strain sensors. The multi-dimensional data records refer to various signal data continuously collected within a specific time period, including high-frequency micro-vibration records reflecting the subtle dynamic changes of the filter under high-frequency vibration conditions, such as vibration amplitude and frequency. In addition, the multi-dimensional data records also include other related multi-dimensional parameter records such as temperature, pressure, and flow rate, which are used to describe the comprehensive state of the environment in which the filter is located. The duration identifier is used to mark the start and end points of these data on the time axis, so that the multi-dimensional data records can be accurately located and analyzed to understand the working state and stress condition of the filter within a specific time period, thereby providing data support for subsequent damage analysis and failure prediction. For example, when the filter operates for one hour, the sensor continuously records the vibration data and corresponding environmental parameters such as temperature rise or pressure change during this period, which together constitute the multi-dimensional data records and have specific time identifiers to ensure the accuracy and traceability of the data.

[0022] Step two: Construct a nonlinear damage accumulation model based on the high-frequency micro-vibration records, multi-dimensional parameter records, and duration identifiers, input the real-time multi-dimensional data into the nonlinear damage accumulation model for vibration cumulative damage analysis, and obtain the high-frequency vibration impact state of the target filter.

[0023] Specifically, based on high-frequency micro-vibration records (referring to vibration data with higher frequency and smaller amplitude captured during the operation of the filter), multi-dimensional parameter records (covering key parameters in the filter operating environment such as temperature, pressure, and flow rate), and duration identifiers (used to mark the specific time period of these data on the time axis), a nonlinear damage accumulation model is constructed by integrating these data. The nonlinear damage accumulation model can handle complex, nonlinear relationship data and is used to evaluate the cumulative damage of materials. During operation, the real-time multi-dimensional data (i.e., vibration, temperature, pressure, and other parameters at the current time) are input into the nonlinear damage accumulation model, and the cumulative effect of the real-time multi-dimensional data during the operation of the filter is analyzed to evaluate the impact state of the filter under high-frequency vibration conditions. For example, the nonlinear damage accumulation model can predict whether the filter will approach failure due to cumulative damage under continuous high-frequency vibration, thereby helping to take timely maintenance measures.

[0024] Step three: Perform risk trend evaluation based on the high-frequency vibration impact state, obtain multi-level risk trends, and perform failure warning based on the multi-level risk trends to obtain multi-level failure warning signals.

[0025] Specifically, the health status of the target filter is evaluated according to the high-frequency vibration influence state (i.e., the degree of influence and cumulative damage of the filter under high-frequency vibration), and the risk change trend of the target filter in the future is analyzed. The risk trend refers to the development of the damage degree and failure risk of the target filter over time. Through risk trend evaluation, the risk is divided into multiple levels (e.g., low, medium, and high risk), forming a multi-level risk trend, which reflects the probability of filter failure under different conditions. Based on the multi-level risk trend, failure warning can be performed, the possible failure time point of the filter in the future can be predicted, and the corresponding multi-level failure warning signal can be generated to prompt the user to take appropriate preventive measures. For example, if a risk trend shows that the damage of the filter is increasing rapidly, a high-level warning signal is issued, suggesting that the filter should be checked or replaced immediately.

[0026] Step four: performing failure prediction according to the multi-level failure warning signal to obtain a failure prediction node.

[0027] Specifically, failure prediction is performed according to the multi-level failure warning signal, i.e., the warning information of different risk levels generated by monitoring high-frequency vibration and other key parameters. Failure prediction refers to predicting the specific time point or condition at which the filter may fail using the multi-level failure warning signal. By analyzing the change trend and intensity of the multi-level failure warning signal, the key moment when the filter gradually approaches failure is identified and defined as a failure prediction node. The failure prediction node is an estimate of the possible failure of the filter at a certain time in the future or under specific working conditions, helping users to take maintenance or replacement measures in advance. For example, if the warning signal shows that high-frequency vibration has increased significantly and exceeded the set threshold, a failure prediction node is generated, indicating that the target filter has a high risk of failure in the future time period.

[0028] Step five: constructing a virtual model of the target filter and performing simulation maintenance on the virtual model to obtain a virtual failure warning node.

[0029] Specifically, constructing a virtual model of the target filter core means creating a digital model corresponding to the actual filter core structure, materials and operating conditions using computer technology. By simulating the working state of the filter core and the various stresses it is subjected to, especially high-frequency vibrations and environmental factors, in a virtual environment, the virtual model can be simulated and maintained. Simulation maintenance is a simulated maintenance operation on the virtual model, such as cleaning or replacing filter material, to evaluate the impact of the simulated maintenance operation on the service life of the filter core. During the simulation, the time point or condition that may lead to failure of the filter core is identified, which is defined as a virtual failure warning node. The virtual failure warning node refers to a key point at which the filter core is predicted to fail at a future time in a simulated environment. For example, if the simulation result shows that the filter core still bears excessive vibration stress for a short time after a certain maintenance, a virtual failure warning node can be determined to prompt the need to further optimize the maintenance strategy.

[0030] Step six: judging the post-effectiveness of the virtual failure warning node on the failure prediction node, and determining whether to perform filter core maintenance according to the judgment result.

[0031] Specifically, judging the post-effectiveness of the virtual failure warning node on the failure prediction node means evaluating the time sequence and relevance between the failure warning node identified in the virtual simulation and the failure prediction node in the actual operation. The virtual failure warning node is a key point of possible failure obtained by simulation, while the failure prediction node is a failure time point obtained based on actual data analysis. If the virtual failure warning node occurs after the failure prediction node, it means that the simulation can effectively prevent the warning failure. Post-effectiveness refers to whether the virtual node is sufficiently late in time to verify the consistency of simulation and actual operation. For example, the prediction in actual operation shows that the filter core fails under a certain high-frequency vibration condition, while in simulation, the filter core fails later under the same condition. By comparing the time of the two, it is judged whether the simulation prediction is accurate, and whether the filter core maintenance needs to be performed in advance is decided accordingly. If the judgment result shows that the simulation is slightly later than the actual failure time point, the operation measure can be selected for preventive maintenance.

[0032] The filter core failure warning method based on multi-dimensional parameter analysis can be applied to a filter core failure warning device based on multi-dimensional parameter analysis, and can achieve the technical goal of comprehensive evaluation and real-time monitoring of filter core multi-dimensional parameters, and achieve the technical effect of accurately predicting the failure time of the filter core and improving the reliability and maintenance efficiency of the equipment.

[0033] Further, the present application also includes:

[0034] According to the analysis of the cumulative damage contribution of the multi-dimensional parameter records to the target filter, the weights of the multi-dimensional parameter records in constructing the nonlinear cumulative damage model are adaptively adjusted, wherein the multi-dimensional parameter records include temperature parameters, pressure parameters, fluid flow pulsation parameters, environmental noise parameters, and equipment operating state parameters.

[0035] Specifically, the analysis of the cumulative damage contribution of the multi-dimensional parameter records to the target filter means determining the contribution of different parameters in the overall damage by evaluating the impact of different parameters on the cumulative damage of the filter. The adaptive adjustment of the weights of the multi-dimensional parameter records in constructing the nonlinear cumulative damage model means dynamically adjusting the importance of each parameter in the model according to the analysis results. The multi-dimensional parameter records include temperature parameters, pressure parameters, fluid flow pulsation parameters, environmental noise parameters, and equipment operating state parameters. The temperature parameters reflect the stress changes of the filter at different temperatures. The pressure parameters record the effects of fluid pressure on the structure of the filter. The fluid flow pulsation parameters involve the impact of fluid flow instability on the filter. The environmental noise parameters evaluate the interference of external noise on vibration data. The equipment operating state parameters include the speed and load conditions of the equipment. Through weight adjustment, the model can more accurately reflect the impact of each parameter on the damage of the filter under actual working conditions. For example, if the impact of the equipment operating state on the damage significantly increases under a certain working condition, the nonlinear cumulative damage model will automatically increase the weight of the equipment operating state parameter to more accurately predict the cumulative damage of the filter.

[0036] By analyzing the cumulative damage contribution of the multi-dimensional parameter records and adaptively adjusting the weights of these parameters, a nonlinear cumulative damage model that reflects the real working conditions of the filter can be more accurately constructed.

[0037] Further, the present application also includes:

[0038] The trend feature extraction of the slope, inflection point, and acceleration of the high-frequency vibration influence state is performed, the risk index value is obtained through risk index calculation according to the trend feature, the multi-level risk grade is obtained through risk threshold division of the risk index value, and the preliminary change trend of the multi-level risk grade is obtained through time series analysis, and the multi-level risk trend is obtained through short-term prediction verification of the preliminary change trend.

[0039] Specifically, the trend feature extraction of the high-frequency vibration influence state refers to extracting features that can reflect the vibration trend by calculating the slope of the vibration signal change, identifying the key turning points of the vibration curve, and measuring the acceleration of the vibration change when analyzing the high-frequency vibration data of the filter element. The slope represents the growth rate of the vibration amplitude over time, the turning point is the time point at which the vibration direction or amplitude changes significantly, and the acceleration reflects the rate of change in vibration intensity. For example, an increasing slope may indicate that the vibration intensity is increasing, a turning point may indicate a change in vibration pattern, and an increase in acceleration may mean that a severe vibration is about to occur. According to the trend feature, the risk index is calculated, which can quantify the risk level of the filter element under different vibration states, thereby obtaining specific risk index values.

[0040] Next, the risk index values are divided into multiple risk levels according to the risk threshold. The risk threshold is a pre-set index value limit used to determine the risk state of the filter element. After comparing the calculated risk index values with these thresholds, the risk can be divided into multiple levels, such as low risk level, medium risk level, and high risk level. Each level represents a different failure probability, helping users understand the current risk level faced by the filter element. For example, when the risk index value exceeds a certain threshold, the state of the filter element may be classified as a high risk level, indicating a high probability of failure.

[0041] The preliminary change trend of the multiple risk levels obtained through time series analysis is to use time series analysis methods to study the evolution of risk levels over time. Time series analysis is a statistical method used to process data arranged in chronological order, by identifying patterns and trends in the data to predict future changes. The preliminary change trend is a preliminary judgment of the future direction of risk levels, such as whether the risk level will increase or decrease over time. Subsequently, short-term prediction verification is performed on the preliminary change trend to further confirm the accuracy of these trends, and ultimately obtain the multi-level risk trend. Short-term prediction verification verifies the consistency between the prediction and the actual result to ensure the reliability of the trend prediction.

[0042] Through trend feature extraction, risk index calculation, risk level division, and time series analysis of the high-frequency vibration influence state, the risk condition of the filter element can be systematically evaluated, and the accuracy of the risk trend can be improved through short-term prediction verification, thereby providing a scientific basis for the maintenance and management of the filter element.

[0043] Further, the present application also includes:

[0044] The multi-level risk trend includes a primary risk trend, an intermediate risk trend, and a high-level risk trend. When the high-frequency micro-vibration data exceeds the primary risk trend and does not reach the intermediate risk trend, a primary early warning signal is sent out. When the high-frequency micro-vibration data exceeds the intermediate risk trend and does not reach the high-level risk trend, an intermediate early warning signal is sent out. When the high-frequency micro-vibration data exceeds the high-level risk trend, a high-level early warning signal is sent out. The primary early warning signal, the intermediate early warning signal, and the high-level early warning signal are integrated to obtain the multi-level failure early warning signal.

[0045] Specifically, the multi-level risk trend includes a primary risk trend, an intermediate risk trend, and a high-level risk trend, which means that the risk level of the filter element is stratified evaluated, and is divided into primary, intermediate, and high-level risks according to different risk levels. The primary risk trend represents a lower risk level, meaning that the filter element is in a normal operating state but may have slight abnormalities. The intermediate risk trend indicates that the risk has increased, and the filter element may have begun to exhibit more significant abnormal conditions. The high-level risk trend represents the highest risk level, and the filter element may be close to failure, necessitating immediate action. For example, over time, the high-frequency vibration data of the filter element may gradually transition from the primary risk trend to the intermediate and high-level risk trends.

[0046] Then, when the high-frequency micro-vibration data exceeds the primary risk trend and does not reach the intermediate risk trend, a primary early warning signal is sent out. The primary early warning signal is a warning sent when the vibration data of the filter element is detected to have just exceeded the primary risk trend, indicating that there may be initial abnormalities, but the situation is not serious. For example, when the vibration amplitude slightly exceeds the normal range but has not reached the intermediate risk level, a primary early warning signal is sent out, reminding the user to pay attention to the status of the filter element and monitor it.

[0047] When the high-frequency micro-vibration data exceeds the intermediate risk trend and does not reach the high-level risk trend, an intermediate early warning signal is sent out. The intermediate early warning signal indicates that the risk level of the filter element has significantly increased, with the vibration data exceeding the intermediate risk trend but not reaching the highest high-level risk trend. For example, when the vibration frequency and amplitude continue to increase, indicating that the health of the filter element is deteriorating, an intermediate early warning signal is sent out, suggesting that an inspection or preliminary maintenance measures be taken.

[0048] Finally, when the high-frequency micro-vibration data exceeds the high-level risk trend, a high-level early warning signal is sent out. The high-level early warning signal is a warning sent when the vibration data is detected to have exceeded the high-level risk trend, indicating that the filter element is in an extremely high risk of failure and may require immediate emergency measures. For example, if the vibration data far exceeds the normal range, indicating that the filter element may be about to fail, a high-level early warning signal is sent out, prompting the user to replace it immediately or stop the operation of the device.

[0049] By monitoring the primary, intermediate and high risk trends, and issuing different levels of early warning signals when the corresponding risk levels exceed the set threshold, the filter failure can be effectively prevented. After integrating the primary, intermediate and high level early warning signals, a multi-level failure early warning signal is formed, providing a comprehensive risk management mechanism to ensure that the filter can be monitored and maintained in time under different risk levels.

[0050] Further, the application also includes:

[0051] The peak vibration, frequency change rate and cumulative vibration energy of the sample multi-level failure early warning signal and the corresponding sample failure time node are extracted, time clustering is performed according to the sample failure time node, and a sample failure probability is generated according to the time clustering result; a filter failure prediction model is constructed based on support vector machine learning of the extracted features, the sample failure time node and the sample failure probability; failure prediction is performed according to the input of the multi-level failure early warning signal into the filter failure prediction model, and a failure probability and a failure prediction time node are obtained; and the failure prediction node is generated according to the failure probability and the failure prediction time node.

[0052] Specifically, the peak vibration, frequency change rate and cumulative vibration energy of the sample multi-level failure early warning signal and the corresponding sample failure time node are extracted, which means that key features are extracted from sample data that have already failed. Peak vibration refers to the maximum value of vibration intensity, frequency change rate reflects the change speed of vibration frequency, and cumulative vibration energy is the total energy of vibration within a certain time. These features can reveal the vibration behavior of the filter at different failure stages. For example, a certain filter may have experienced rapid frequency increase and dramatic cumulative energy increase before failure. According to these characteristic data, combined with the sample failure time node, time clustering can be performed, i.e. samples with similar failure times are classified into one category to identify time patterns and potential failure rules, thereby generating a sample failure probability representing the possibility of failure within a certain time.

[0053] Then, based on support vector machine learning of the extracted features, sample failure time node and sample failure probability, a filter failure prediction model is constructed. Support vector machine is a supervised learning model specifically used for classification and regression analysis. In this application, support vector machine learns the extracted vibration features, sample failure time node and corresponding failure probability to identify the relationship between features and failure. Through this learning process, support vector machine generates a prediction model to judge the failure probability of new input data. For example, by identifying similar vibration patterns and time nodes, support vector machine can predict the failure risk of the filter at a certain future time point.

[0054] Then, the multi-stage failure early warning signals are input to the filter failure prediction model to perform failure prediction, and failure probability and failure prediction time node are obtained. The multi-stage failure early warning signals contain vibration information of different risk levels. After inputting these signals to the previously constructed failure prediction model, the model outputs the probability of filter failure and the prediction time node under the current condition. The failure probability represents the possibility of filter failure under the current operating condition, while the failure prediction time node is the specific time point predicted by the model at which the filter may fail. For example, if the current vibration signal shows high frequency and high energy accumulation, the model may predict a high failure probability and a near failure time node.

[0055] Finally, a failure prediction node is generated according to the failure probability and the failure prediction time node. The failure prediction node refers to a specific time point at which the filter has a higher possibility of failure. By combining the failure probability and the time node output by the model, a specific time range can be determined as the failure prediction node. For example, if the model predicts that a filter has an 80% probability of failure within the next week, the end of the week may be marked as the failure prediction node.

[0056] Through feature extraction, time clustering and support vector machine learning, a filter failure prediction model can be effectively constructed. Based on the model, by inputting multi-stage failure early warning signals, the failure probability and time node of the filter can be accurately predicted, and finally the failure prediction node is generated, providing data support for preventive maintenance.

[0057] Further, the present application also includes:

[0058] The stress distribution, deformation condition and material fatigue of the virtual model are evaluated to obtain a static simulation result; cumulative damage of high-frequency micro-vibration is configured to the static simulation result to obtain an initial virtual failure early warning node of an initial simulation result; the stress distribution, deformation condition and material fatigue of the initial simulation result after performing a simulation maintenance operation are evaluated to obtain a maintenance effect, and the virtual failure early warning node based on the maintenance effect is calculated.

[0059] Specifically, evaluating the stress distribution, deformation condition and material fatigue of the virtual model to obtain a static simulation result means that in a simulated environment, the mechanical stress, structural deformation and material fatigue effect of the filter under normal operating conditions are determined through calculation and analysis. Stress distribution describes the magnitude and distribution of forces borne by different parts of the filter, deformation condition refers to the deformation of the filter under stress, and material fatigue reflects the damage that may occur to the material under repeated loading. For example, under static conditions, some parts of the filter may bear a large stress, causing slight structural deformation in that part.

[0060] Next, the initial simulation results are configured with high-frequency micro-vibration cumulative damage to obtain the initial virtual failure warning node. High-frequency micro-vibration cumulative damage refers to the gradual micro-damage caused by the filter under the influence of cumulative stress in a high-frequency vibration environment. By introducing high-frequency vibration factors based on static simulation, the damage accumulation process of the filter under actual working conditions is simulated. The initial virtual failure warning node is the time point at which the first possible failure is predicted by simulation analysis, indicating that the filter material may be close to failure due to cumulative damage at this time point.

[0061] Then, the stress distribution, deformation, and material fatigue of the initial simulation results after performing simulated maintenance operations are evaluated to obtain the maintenance effect, and the virtual failure warning node based on the maintenance effect is calculated. Simulated maintenance operations refer to a series of simulated operations in the virtual model, such as cleaning or replacing filter materials, to evaluate the impact of maintenance measures on filter performance. The results after evaluation reflect the stress distribution and material state of the filter after maintenance, as well as the impact of these changes on the service life of the filter. The virtual failure warning node is recalculated and adjusted to determine the time point at which the filter may fail under new operating conditions after maintenance. For example, by simulating cleaning operations, the stress on certain parts of the filter can be reduced, thereby extending its service life.

[0062] Through static simulation, high-frequency vibration cumulative damage analysis, and simulated maintenance operations, the state and life of the filter under different working conditions can be comprehensively evaluated, and the possible failure time of the filter can be predicted through the calculation of the virtual failure warning node, guiding the actual maintenance and replacement decisions.

[0063] Further, the present application also includes:

[0064] The virtual failure warning node is compared with the failure prediction node in time series analysis to determine the post-effectiveness of the virtual failure warning node on the failure prediction node. If the result of the determination is post-invalidation, an extended search of the simulation maintenance strategy is performed, and if the result of the determination is post-effectiveness, the filter is maintained.

[0065] Specifically, comparing the virtual failure warning node with the failure prediction node in time series analysis means analyzing the order and corresponding relationship of the two in time to evaluate the matching degree between the failure node predicted by the virtual model and the failure node inferred from actual data. Time series analysis is a method for comparing two sets of time-related data to determine their correlation and time difference. The virtual failure warning node is the expected failure time point obtained by simulation, while the failure prediction node is the failure time inferred based on actual operation data. By comparing, the prediction accuracy and reliability of the virtual simulation model can be evaluated.

[0066] Next, the post-position validity of the virtual failure warning node to the failure prediction node is determined, that is, whether the virtual failure warning node is reasonably located after the failure prediction node in time is analyzed. The post-position validity means that the virtual failure warning node is after the actual failure time. If the virtual node time is after the actual failure time, it means that the virtual maintenance operation is effective and delays the failure time node. If the judgment result is post-position invalid, the expansion search of the simulation maintenance strategy is performed, that is, more maintenance strategies or adjustment of existing strategies are explored in the virtual model to improve the accuracy and practicality of the simulation result. The expansion search refers to finding a more suitable maintenance method by simulating different maintenance scenarios or adjusting parameters, the purpose being to optimize the position of the virtual failure warning node to make it more consistent with the actual situation. For example, by increasing the number of simulated maintenance operations or adjusting the maintenance time, the prediction of the simulation model can be made more close to the actual failure condition.

[0067] If the judgment result is post-position valid, the filter core maintenance is performed, which means that the time point of the virtual failure warning node is reasonable and reliable and can be used as a reference basis for actual maintenance. The judgment result of post-position valid indicates that the prediction of the virtual model is consistent with the failure time point under the actual operating condition, so the actual filter core maintenance operation such as replacing the filter core or other preventive maintenance can be performed according to the time of the virtual failure warning node.

[0068] By comparing and analyzing the virtual failure warning node and the failure prediction node through time series, the prediction accuracy of the simulation model can be evaluated, and appropriate maintenance measures or optimization of the simulation strategy can be taken according to the judgment result, so as to ensure the reliable operation of the filter core and prolong the service life.

[0069] In summary, the filter core failure warning method based on multi-dimensional parameter analysis provided in the present application has the following technical effects:

[0070] The multi-dimensional data record of the target filter element and real-time multi-dimensional data are obtained through interactive multi-modal sensors, the multi-dimensional data record includes high-frequency micro-vibration record and corresponding multi-dimensional parameter record, the multi-dimensional parameter record has a time duration identifier; a nonlinear damage accumulation model is constructed according to the high-frequency micro-vibration record, the multi-dimensional parameter record and the time duration identifier, vibration cumulative damage analysis is performed based on the real-time multi-dimensional data input into the nonlinear damage accumulation model, and a high-frequency vibration influence state of the target filter element is obtained; risk trend assessment is performed according to the high-frequency vibration influence state, a multi-level risk trend is obtained, and multi-level failure warning signals are obtained based on the multi-level risk trend; failure prediction is performed according to the multi-level failure warning signals, and a failure prediction node is obtained; a virtual model of the target filter element is constructed, and a virtual failure warning node is obtained by simulating and maintaining the virtual model; the post-effectiveness of the virtual failure warning node on the failure prediction node is judged, and it is determined whether to perform filter element maintenance according to the judgment result, the technical target of multi-dimensional parameter comprehensive evaluation and real-time monitoring of the filter element is achieved, and the technical effect of accurately predicting the failure time of the filter element and improving the operation reliability and maintenance efficiency of the equipment is achieved.

[0071] Embodiment two, based on the filter element failure warning method based on multi-dimensional parameter analysis in the foregoing embodiments, the same invention concept, the present application also provides a filter element failure warning device based on multi-dimensional parameter analysis, please refer to the attached Figure 2 , including:

[0072] The multi-dimensional data record obtaining module 11 is used for obtaining multi-dimensional data records and real-time multi-dimensional data of the target filter element by an interactive multi-modal sensor, the multi-dimensional data records include high-frequency micro-vibration records and corresponding multi-dimensional parameter records, and the multi-dimensional parameter records have time duration identifiers; the high-frequency vibration influence state obtaining module 12 is used for constructing a non-linear damage accumulation model according to the high-frequency micro-vibration records, the multi-dimensional parameter records and the time duration identifiers, performing vibration cumulative damage analysis based on the real-time multi-dimensional data input into the non-linear damage accumulation model, and obtaining a high-frequency vibration influence state of the target filter element; the multi-level failure warning signal obtaining module 13 is used for performing risk trend evaluation according to the high-frequency vibration influence state, obtaining a multi-level risk trend, and performing failure warning based on the multi-level risk trend to obtain a multi-level failure warning signal; the failure prediction node obtaining module 14 is used for performing failure prediction according to the multi-level failure warning signal to obtain a failure prediction node; the virtual failure warning node obtaining module 15 is used for constructing a virtual model of the target filter element, performing simulation maintenance on the virtual model to obtain a virtual failure warning node; and the filter element maintenance module 16 is used for judging the post-position effectiveness of the virtual failure warning node to the failure prediction node, and determining whether to perform filter element maintenance according to a judgment result.

[0073] Further, the filter element failure warning device based on multi-dimensional parameter analysis is also used for:

[0074] According to damage accumulation contribution analysis of the target filter element based on the multi-dimensional parameter records, the weight of the multi-dimensional parameter records in constructing the non-linear damage accumulation model is adaptively adjusted, wherein the multi-dimensional parameter records include temperature parameters, pressure parameters, fluid flow pulsation parameters, environmental noise parameters and equipment running state parameters.

[0075] Further, the filter element failure warning device based on multi-dimensional parameter analysis is also used for:

[0076] Trend feature extraction of a slope, an inflection point and an acceleration is performed on the high-frequency vibration influence state, risk index calculation is performed according to the trend features to obtain a risk index value, multi-level risk grades are obtained by dividing the risk index value according to a risk threshold value, a preliminary change trend of the multi-level risk grades is obtained through time series analysis, short-term prediction verification is performed on the preliminary change trend, and the multi-level risk trend is obtained.

[0077] Further, the filter element failure warning device based on multi-dimensional parameter analysis is also used for:

[0078] The multi-level risk trend comprises a primary risk trend, an intermediate risk trend and a high-level risk trend; a primary early warning signal is sent when the high-frequency micro-vibration data exceeds the primary risk trend and does not reach the intermediate risk trend; an intermediate early warning signal is sent when the high-frequency micro-vibration data exceeds the intermediate risk trend and does not reach the high-level risk trend; a high-level early warning signal is sent when the high-frequency micro-vibration data exceeds the high-level risk trend; and the primary early warning signal, the intermediate early warning signal and the high-level early warning signal are integrated to obtain the multi-level failure early warning signal.

[0079] Further, the filter element failure early warning device based on multi-dimensional parameter analysis is further used for:

[0080] The peak vibration, the frequency change rate and the cumulative vibration energy of the sample multi-level failure early warning signal and the corresponding sample failure time node are extracted, time clustering is performed according to the sample failure time node, and a sample failure probability is generated according to the time clustering result; a support vector machine is used to learn the extracted features, the sample failure time node and the sample failure probability, and a filter element failure prediction model is constructed; failure prediction is performed according to the multi-level failure early warning signal input into the filter element failure prediction model, and a failure probability and a failure prediction time node are obtained; and the failure prediction node is generated according to the failure probability and the failure prediction time node.

[0081] Further, the filter element failure early warning device based on multi-dimensional parameter analysis is further used for:

[0082] The stress distribution, deformation and material fatigue of the virtual model are evaluated to obtain a static simulation result; the static simulation result is configured with cumulative damage of high-frequency micro-vibration to obtain an initial virtual failure early warning node of an initial simulation result; the stress distribution, deformation and material fatigue of the initial simulation result after performing a simulation maintenance operation are evaluated to obtain a maintenance effect, and the virtual failure early warning node based on the maintenance effect is calculated.

[0083] Further, the filter element failure early warning device based on multi-dimensional parameter analysis is further used for:

[0084] The virtual failure early warning node and the failure prediction node are compared and analyzed in time sequence to determine the post-position effectiveness of the virtual failure early warning node on the failure prediction node; if the determination result is post-position invalid, simulation maintenance strategy expansion search is performed, and if the determination result is post-position valid, filter element maintenance is performed.

[0085] The various embodiments are described in the specification, each of which focuses on the differences from other embodiments, the filter core failure early warning method based on multi-dimensional parameter analysis in the foregoing embodiment one and the specific examples are also applicable to the filter core failure early warning device based on multi-dimensional parameter analysis in the present embodiment, through the foregoing detailed description of the filter core failure early warning method based on multi-dimensional parameter analysis, those skilled in the art can clearly know the filter core failure early warning device based on multi-dimensional parameter analysis in the present embodiment, therefore, for the sake of brevity of the specification, it will not be described in detail here. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant part is described in the method part.

[0086] The above description of disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0087] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and 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 equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A filter cartridge failure warning method based on multi-dimensional parameter analysis, characterized in that, The method comprises the following steps: An interactive multi-modal sensor obtains multi-dimensional data records and real-time multi-dimensional data of a target filter element, the multi-dimensional data records include high-frequency micro-vibration records and corresponding multi-dimensional parameter records, and the multi-dimensional parameter records have time duration identifiers; A non-linear damage accumulation model is constructed according to the high-frequency micro-vibration records, the multi-dimensional parameter records and the time duration identifiers, vibration cumulative damage analysis is performed based on the real-time multi-dimensional data input into the non-linear damage accumulation model, and a high-frequency vibration influence state of the target filter element is obtained; Risk trend assessment is performed according to the high-frequency vibration influence state, multi-level risk trends are obtained, and multi-level failure warning signals are obtained based on the multi-level risk trends; Failure prediction is performed according to the multi-level failure warning signals, and a failure prediction node is obtained, comprising: Characteristic extraction of peak vibration, frequency change rate and cumulative vibration energy is performed on sample multi-level failure warning signals and corresponding sample failure time nodes, time clustering is performed according to the sample failure time nodes, and a sample failure probability is generated according to the time clustering result; Support vector machine is used to learn the extracted characteristics, the sample failure time nodes and the sample failure probability, and a filter element failure prediction model is constructed; Failure prediction is performed according to input of the multi-level failure warning signals into the filter element failure prediction model, and a failure probability and a failure prediction time node are obtained; The failure prediction node is generated according to the failure probability and the failure prediction time node; A virtual model of the target filter element is constructed, and a virtual failure warning node is obtained by simulating maintenance of the virtual model, comprising: Stress distribution, deformation and material fatigue of the virtual model are evaluated, and a static simulation result is obtained; Cumulative damage of high-frequency micro-vibration is configured to the static simulation result, and an initial virtual failure warning node of an initial simulation result is obtained; Stress distribution, deformation and material fatigue of the initial simulation result after performing a simulated maintenance operation are evaluated, a maintenance effect is obtained, and the virtual failure warning node based on the maintenance effect is calculated; The post-position effectiveness of the virtual failure warning node on the failure prediction node is judged, and it is determined whether to perform filter element maintenance according to the judgment result.

2. The filter cartridge failure warning method based on multi-dimensional parameter analysis according to claim 1, wherein, Damage accumulation contribution analysis of the target filter element is performed according to the multi-dimensional parameter records, and the weights of the multi-dimensional parameter records in constructing the non-linear damage accumulation model are adaptively adjusted, wherein the multi-dimensional parameter records include temperature parameters, pressure parameters, fluid flow pulsation parameters, environmental noise parameters and equipment operating state parameters.

3. The filter cartridge failure warning method based on multi-dimensional parameter analysis according to claim 1, characterized in that, The multi-level risk trends are obtained, comprising: Trend characteristic extraction of slope, inflection point and acceleration is performed on the high-frequency vibration influence state, risk index values are obtained by performing risk index calculation according to the trend characteristics; Multi-level risk grades are obtained by dividing the risk index values according to a risk threshold value; Preliminary change trends of the multi-level risk grades are obtained through time series analysis, short-term prediction verification is performed on the preliminary change trends, and the multi-level risk trends are obtained.

4. The filter cartridge failure warning method based on multi-dimensional parameter analysis of claim 1, wherein, The multi-level failure warning signals are obtained, comprising: The multi-level risk trends include primary risk trends, intermediate risk trends and high-level risk trends; a primary early warning signal is sent when the high-frequency micro-vibration data exceeds the primary risk trend and does not reach the intermediate risk trend; an intermediate early warning signal is sent when the high-frequency micro-vibration data exceeds the intermediate risk trend and does not reach the advanced risk trend; an advanced early warning signal is sent when the high-frequency micro-vibration data exceeds the advanced risk trend; the primary early warning signal, the intermediate early warning signal and the advanced early warning signal are integrated to obtain the multi-level failure early warning signal.

5. The filter cartridge failure warning method based on multi-dimensional parameter analysis of claim 1, wherein, The judgment of the post-position effectiveness of the virtual failure early warning node to the failure prediction node and the determination of whether to perform filter maintenance according to the judgment result include: The time sequence comparison analysis of the virtual failure early warning node and the failure prediction node is performed to judge the post-position effectiveness of the virtual failure early warning node to the failure prediction node. If the judgment result is post-position invalid, the expansion search of the simulation maintenance strategy is performed, and if the judgment result is post-position valid, the filter maintenance is performed.

6. The filter cartridge failure warning device based on multi-dimensional parameter analysis, characterized in that, The steps for implementing the filter failure early warning method based on multi-dimensional parameter analysis in any one of claims 1 to 5 include: A multi-dimensional data record obtaining module is used to obtain multi-dimensional data records and real-time multi-dimensional data of a target filter through interactive multi-modal sensors, the multi-dimensional data records include high-frequency micro-vibration records and corresponding multi-dimensional parameter records, and the multi-dimensional parameter records have time duration identifiers; A high-frequency vibration influence state obtaining module is used to construct a non-linear damage accumulation model according to the high-frequency micro-vibration records, the multi-dimensional parameter records and the time duration identifiers, perform vibration cumulative damage analysis based on the real-time multi-dimensional data input into the non-linear damage accumulation model, and obtain the high-frequency vibration influence state of the target filter; A multi-level failure early warning signal obtaining module is used to perform risk trend evaluation according to the high-frequency vibration influence state, obtain multi-level risk trends, perform failure early warning based on the multi-level risk trends, and obtain multi-level failure early warning signals; A failure prediction node obtaining module is used to perform failure prediction according to the multi-level failure early warning signals and obtain failure prediction nodes; A virtual failure early warning node obtaining module is used to construct a virtual model of a target filter, perform simulation maintenance on the virtual model, and obtain virtual failure early warning nodes; A filter maintenance module is used to judge the post-position effectiveness of the virtual failure early warning node to the failure prediction node and determine whether to perform filter maintenance according to the judgment result.

Images