A device fault prediction method and device based on a digital twin model

Abstract

This application relates to a device fault prediction method and device based on a digital twin model. The method includes: when the device monitoring twin model fails to detect a device fault, first obtain the operating status, temperature, and load cycle data of the device. Substitute the operating status data into the fixed fault and usage fault calculation models to obtain the corresponding fixed fault data and usage fault data. Subsequently, correct the fixed fault data according to the temperature data to obtain the temperature-corrected fixed fault data; correct the usage fault data according to the load cycle data to obtain the load-corrected usage fault data. Then, combine the temperature-corrected fixed fault data and the load-corrected usage fault data to calculate the failure data of the device on the usage time axis, and further obtain the failure curve of the device. Finally, extract the data points that meet the preset failure conditions from the failure curve as the device fault prediction information, which can effectively improve the real-time performance and accuracy in the prediction of device faults.

Description

Technical Field

[0001] The present application relates to the field of computer technologies, and in particular, to a method and device for predicting equipment failures based on a digital twin model. Background Art

[0002] In traditional technologies, equipment failure prediction usually involves monitoring the operating parameters of equipment (such as temperature, pressure, vibration, etc.), comparing them with historical data, and establishing a normal operating mode for the equipment. Once a deviation in the parameters is detected, a warning mechanism is triggered. In addition, traditional methods also include experience-based failure mode analysis (such as Failure Mode and Effects Analysis, FMEA), periodic inspections, and artificial expert systems. By means of regular inspections, data statistics, and empirical rules, combined with the service life and historical failure data of the equipment, potential failure points are predicted, thereby achieving the maintenance and preventive maintenance of the equipment. However, traditional technologies often rely on empirical rules and manual inspections in equipment failure prediction, making it difficult to handle complex and changing system environments, resulting in a lack of real-time performance and accuracy in equipment failure prediction. Summary of the Invention

[0003] Based on this, in view of the above technical problems, it is necessary to provide a method, device, computer device, computer-readable storage medium, and computer program product for predicting equipment failures based on a digital twin model, which can effectively improve the real-time performance and accuracy in equipment failure prediction.

[0004] In a first aspect, the present application provides a method for predicting equipment failures based on a digital twin model, including:

[0005] In the case where the equipment monitoring twin model fails to detect fault information of the target equipment, obtaining the equipment operating status data, equipment operating temperature data, and equipment load cycle data of the target equipment;

[0006] Substituting the equipment operating status data into the fixed fault calculation model and the usage fault calculation model of the target equipment respectively to obtain target fixed fault data and target usage fault data;

[0007] Correcting the target fixed fault data according to the equipment operating temperature data to obtain temperature-corrected fixed fault data;

[0008] Correcting the target usage fault data according to the equipment load cycle data to obtain load-corrected usage fault data;

[0009] Calculating the failure data of the target equipment on the usage time axis according to the temperature-corrected fixed fault data and the load-corrected usage fault data to obtain an equipment failure curve;

[0010] Extract data points that meet the preset failure conditions from the device failure curve as the device fault prediction information of the target device.

[0011] In a second aspect, the present application also provides a device fault prediction device based on a digital twin model, including:

[0012] A device data acquisition module, configured to acquire the device operation status data, device operation temperature data, and device load cycle data of the target device when the device monitoring twin model fails to detect fault information of the target device;

[0013] A fault data calculation module, configured to substitute the device operation status data into the fixed fault calculation model and the usage fault calculation model of the target device respectively to obtain the target fixed fault data and the target usage fault data;

[0014] A fault data correction module, configured to correct the target fixed fault data according to the device operation temperature data to obtain temperature-corrected fixed fault data;

[0015] The fault data correction module is further configured to correct the target usage fault data according to the device load cycle data to obtain load-corrected usage fault data;

[0016] A failure curve generation module, configured to calculate the failure data of the target device on the usage time axis according to the temperature-corrected fixed fault data and the load-corrected usage fault data to obtain a device failure curve;

[0017] A device fault prediction module, configured to extract data points that meet the preset failure conditions from the device failure curve as the device fault prediction information of the target device.

[0018] In a third aspect, the present application also provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, any step of a device fault prediction method based on a digital twin model is implemented.

[0019] The above device fault prediction method, device and computer device based on the digital twin model input the device operation state data into the fixed fault calculation model and the usage fault calculation model respectively to generate the object fixed fault data and object usage fault data of the target device. Then, the temperature data of the device is used to correct the fixed fault data, and the load cycle data is used to correct the usage fault data, so as to more accurately reflect the failure trend of the device under different working environments. Further, the failure curve of the device is calculated by using the two corrected fault data, and the fault prediction information meeting the preset failure conditions is extracted from it to realize the early warning of device faults. Through the above fault prediction mechanism, the real-time performance and accuracy in the prediction of device faults can be effectively improved, the reliability of device operation can be further significantly improved, the maintenance and repair cycle can be optimized, the downtime caused by sudden faults can be reduced, the maintenance cost can be lowered, and the overall operation and maintenance efficiency of the device can be improved, thus providing a strong guarantee for the long-term stable operation of the device. Brief Description of the Drawings

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 It is an application environment diagram of the device fault prediction method based on the digital twin model in an embodiment;

[0022] Figure 2 It is a schematic flowchart of the device fault prediction method based on the digital twin model in an embodiment;

[0023] Figure 3 It is a schematic flowchart of the method for obtaining the temperature-corrected fixed fault data in an embodiment;

[0024] Figure 4 It is a schematic flowchart of the method for obtaining the load-corrected usage fault data in an embodiment;

[0025] Figure 5 It is a schematic flowchart of the method for obtaining the first device failure curve in an embodiment;

[0026] Figure 6 It is a schematic flowchart of the method for obtaining the second device failure curve in an embodiment;

[0027] Figure 7 It is a schematic flowchart of the method for obtaining the third device failure curve in an embodiment;

[0028] Figure 8 It is a structural block diagram of a device fault prediction device based on a digital twin model in an embodiment;

[0029] Figure 9 It is an internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0030] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0031] A device fault prediction method based on a digital twin model provided by an embodiment of the present application can be applied to, for example, Figure 1 the application environment shown in the figure. Among them, the terminal 101 communicates with the server 102 through the network. The data storage system can store the data that the server 102 needs to process. The data storage system can be integrated on the server 102, or can be placed in the cloud or other network servers. Among them, the server 102 can be implemented by an independent server or a server cluster composed of multiple servers.

[0032] In an exemplary embodiment, as Figure 2 shown in the figure, a device fault prediction method based on a digital twin model is provided. Taking the method applied to Figure 1 the server in the figure as an example for description, it includes the following steps 201 to step 206. Among them:

[0033] Step 201, in the case that the device monitoring twin model fails to detect fault information of the target device, obtain the device operation state data, device operation temperature data, and device load cycle data of the target device.

[0034] Among them, the device monitoring twin model can be a virtual simulation model of the target device, which generates a corresponding digital model by obtaining the data of the physical device in real time, and simulates the operation state and behavior of the device.

[0035] Among them, the target device can be a specific device or mechanical system that is studied, analyzed, or repaired during the monitoring and prediction process.

[0036] Among them, the fault information can be any abnormal state or performance degradation indication data that may occur during the operation of the device, and usually includes the fault type, occurrence time, influence range of the device, and its impact on production or operation.

[0037] Among them, the device operation status data can be various real-time data describing the current working conditions of the device, including parameters such as the running speed, torque, power, vibration, and pressure of the device.

[0038] Among them, the device operation temperature data can be the temperature data recorded in real time for each component of the device during operation.

[0039] Among them, the device load cycle data can be the record of the load change over time during the operation of the device, including the load intensity, frequency, and change pattern borne by the device in different time periods.

[0040] Specifically, in the case where the device monitoring twin model fails to detect fault information of the target device, the real-time operation status data, device operation temperature data, and load cycle data of the target device are obtained through the data acquisition system. Among them, the real-time operation status data reflects the current working conditions of the device, the device operation temperature data provides the thermal status of the device in the working environment or the working environment in a previous period of time, and the device load cycle data reflects the operation cycle characteristics of the device under different load conditions.

[0041] Step 202, substitute the device operation status data into the fixed fault calculation model and the usage fault calculation model of the target device respectively to obtain the target fixed fault data and the target usage fault data.

[0042] Among them, the fixed fault calculation model can be used to predict the faults that occur in the device due to factors such as material aging, fatigue, and corrosion during long-term operation.

[0043] Among them, the usage fault calculation model can be a fault occurrence model that predicts the device under actual working loads and operating conditions.

[0044] Among them, the target fixed fault data can be the fault data obtained by substituting the device operation status data into the fixed fault calculation model, which represents the potential fault risk of the device caused only by fixed factors such as material aging and fatigue within a specific time period.

[0045] Among them, the target usage fault data can be the fault data obtained by substituting the device operation status data into the usage fault calculation model, which reflects the influence of the usage conditions of the device in actual operation on the occurrence of faults, including the fault risks caused by factors such as overload, long-term high-load operation, and frequent start-stop.

[0046] Specifically, substitute the device operation status data of the device (such as operation duration, load condition, vibration, rotation speed, etc.) into the fixed-fault calculation model. The fixed-fault calculation model assumes that device components will gradually deteriorate due to factors such as material aging, fatigue, and corrosion during long-term operation, generating fixed-fault data of the device. This part of the faults is usually not affected by the usage environment and is more related to time and the degree of device aging. At the same time, substitute the device operation status data of the device (such as operation duration, load condition, vibration, rotation speed, etc.) into the usage-fault calculation model. The usage-fault calculation model takes into account the working conditions of the device under different loads and environments and can estimate the usage-fault data caused by the operating conditions of the device (such as overload, frequent start-stop, etc.). These faults are usually closely related to the working intensity, usage frequency, and operation mode of the device. Through the calculations of these two models, the object fixed-fault data and the object usage-fault data can be obtained respectively.

[0047] Step 203: Correct the object fixed-fault data according to the device operation temperature data to obtain the temperature-corrected fixed-fault data.

[0048] Among them, the temperature-corrected fixed-fault data can correct the fault data of the device by considering the operating characteristics of the device at different temperatures. Especially in a high-temperature environment, the aging and fatigue processes of the device usually accelerate. Therefore, through temperature correction, the fault risk of the device in the actual temperature environment can be more accurately reflected.

[0049] Specifically, during the fault prediction process of the device, the operating temperature of the device is an important factor because a high-temperature environment will accelerate the aging and damage of device components, especially phenomena such as thermal fatigue, material expansion, and chemical reactions. Therefore, the Arrhenius model is used to quantify the impact of the device operating temperature on the fault occurrence rate. The Arrhenius model is based on the relationship between temperature and the chemical reaction rate and assumes that for every certain increase in temperature, the fault occurrence rate of the device will increase exponentially. In specific implementation, the actual operating temperature data of the device will be used as input and passed to the temperature factor in the Arrhenius model. By calculating the impact of temperature on the fault rate, the fixed fault data of the device is then adjusted. That is, the operating state of the device at different temperatures may cause changes in the physical and chemical properties of the material. For example, metal materials may accelerate lattice fatigue in a high-temperature environment, plastic materials may become soft or aged, and lubricating oil may also degrade at high temperatures, thus affecting the life and reliability of device components. Therefore, the Arrhenius model will correct based on the current working temperature of the device according to the impact of temperature on the fault rate. When the device is in a high-temperature environment, the model will increase the probability of fault occurrence, reflecting the accelerated damage caused by high temperature; conversely, when the temperature is low, the fault rate will decrease accordingly. By correcting the fixed fault data with this Arrhenius model, it can accurately reflect the actual occurrence of device faults under different temperature conditions, and obtain temperature-corrected fixed fault data.

[0050] Step 204, correct the object usage fault data according to the device load cycle data to obtain load-corrected usage fault data.

[0051] Among them, the load-corrected usage fault data can be obtained by substituting the working data of the device under different load cycle conditions into the load correction model to adjust the original fault data, reflecting the actual fault risk of the device in working environments such as high load and frequent load changes. This correction helps to accurately predict the occurrence of device faults under complex usage conditions.

[0052] Specifically, device load cycle data is an important factor in device usage failures because the device often experiences varying degrees of load changes during actual operation, which can affect its fatigue life and failure modes. When using device load cycle data to correct usage failure data, the Coffin–Manson model is adopted to consider the impact of load on the device's fatigue behavior. The Coffin–Manson model is mainly used to describe the process of fatigue damage and crack propagation of materials under alternating load conditions. Specifically, the Coffin–Manson model calculates the fatigue life and the acceleration degree of crack propagation by considering factors such as the amplitude and frequency of the load-corrected usage failure data. The core of the model is to establish the failure occurrence law of the device under load conditions based on the relationship between the load-corrected usage failure data and the plastic deformation of the material. When the load cycle of the device increases, the fatigue accumulation of the material and the development speed of microcracks will accelerate, resulting in earlier failures of the device. Therefore, by inputting the device's load cycle data into the Coffin–Manson model, the usage failure data can be accurately corrected, reflecting the likelihood of device failures under working conditions such as high load, frequent start-stop, or vibration. The corrected data more realistically reflects the fatigue damage process of the device under different loads, helps to improve the accuracy of failure prediction, and provides a scientific basis for the maintenance and management of the device.

[0053] During the actual operation of the device, it will experience different load cycles, generating device load cycle data. The amplitude and frequency of the device load cycle data directly affect the fatigue damage of the device components. To accurately evaluate the fault risk of the device under different load conditions, the Coffin–Manson model is used to correct the usage fault data. Starting from the fatigue mechanism of the material, this model believes that when the device is subjected to alternating loads, plastic deformation and the cumulative expansion of microcracks will occur. Especially under high loads or frequent load changes, the fatigue damage of the device components will accelerate. Therefore, the Coffin–Manson model not only considers the amplitude of the load (the intensity of the load change), but also the frequency of the load (the rate of the load change). Because the greater the load amplitude, the more severe the plastic deformation and crack expansion of the component, and the faster the accumulation rate of fatigue damage; the higher the load frequency, the more load cycles the device experiences per unit time, thus exacerbating the fatigue process, leading to crack expansion and premature occurrence of faults. Specifically in the implementation process, by collecting the device load cycle data of the device, analyzing the amplitude, frequency, and number of cycles of its load, and then substituting these data into the Coffin–Manson model, the model will calculate the fatigue life of the device components and the acceleration degree of crack expansion based on these load characteristics, reflecting the specific development of the device fatigue damage under different working load conditions. Furthermore, it corrects the original object usage fault data to obtain temperature-corrected fixed fault data, making the prediction of the impact of the load on the device fault more accurate. Especially in scenarios with large load fluctuations or long-term high-load operation, it can identify potential fatigue fault risks of the device in advance, optimizing the device maintenance cycle and fault warning mechanism.

[0054] In a specific embodiment, assume that a device is operating in a certain factory. The device load cycle data shows that when the device is operating under high loads, the load amplitude reaches the maximum value, and there are 20 load changes per hour (i.e., a relatively high load frequency). Based on these data, first record the actual load conditions of the device, such as the range of fluctuations and the change frequency of each load. Then, input these load data into the Coffin–Manson model. The model will calculate the impact of each load cycle on the plastic deformation and fatigue accumulation of the device components based on different combinations of the load amplitude and frequency. For example, if the device operates under high loads for a long time, the model may show an accelerated crack expansion rate due to the high load amplitude, and the relatively high load frequency further exacerbates the accumulation of fatigue damage. Finally, the model corrects the object usage fault data according to these calculation results, making the predicted device fault time more consistent with the actual load conditions.

[0055] Step 205: Calculate the failure data of the target device on the usage time axis based on the temperature-corrected fixed fault data and the load-corrected usage fault data to obtain the device failure curve.

[0056] Among them, the usage time axis can be the cumulative time data of all time periods passed by the device from the time of commissioning to the preset usage time.

[0057] Among them, the failure data can be the probability of the device possibly having a fault predicted according to the calculation model, and corresponding time, frequency, and type are matched to each probability data.

[0058] Among them, the device failure curve can be a graphical representation of the probability of the device having a fault at different time points or usage conditions. It usually uses time or usage cycle as the horizontal axis and the fault probability or failure rate of the device as the vertical axis, depicting the trend and possibility of the device having a fault during the entire usage process as time goes by or the usage load changes, and marking the corresponding time, frequency, and type at each data point on the vertical axis.

[0059] Specifically, based on the temperature-corrected fixed fault data, the temperature-corrected fixed fault data is integrated respectively with each usage time data point on the usage time axis of the device as the upper limit of integration, and the integration results of all usage time data points are used to establish the failure probability curve of the fixed fault; this curve reflects the probability that the device may have a fault as the usage time increases under different temperature environments. Similarly, the usage fault data is corrected using the load, and the temperature-corrected fixed fault data is integrated respectively with each usage time data point on the usage time axis of the device as the upper limit of integration to calculate the failure probability curve of the usage fault, indicating the probability of the device having a fault under different load conditions. These two curves respectively represent the failure behaviors of the device under the influence of two different factors, temperature and load, but they are calculated separately based on different corrected data. In order to obtain a more comprehensive failure prediction, next, these two failure probability curves are synthesized through the curve superposition method. Specifically, the superposition method synthesizes the failure probabilities of these two curves on the time axis, comprehensively considering the combined influence of temperature and load on the device fault, so as to obtain a comprehensive failure probability curve.

[0060] Step 206, extract the data points that meet the preset failure conditions from the device failure curve as the device fault prediction information of the target device.

[0061] Among them, the preset failure conditions can be specific criteria set according to the device operation characteristics, historical fault data, or factory preset conditions for judging whether the device is approaching a fault.

[0062] Among them, the device fault prediction information can be the prediction result generated by analyzing the device operation data, environmental impact, and corrected fault data, including the time node when the device may have a fault in the future, the fault type, and the probability of the fault occurring.

[0063] Specifically, preset failure conditions are set according to the target device, such as the critical probability of equipment failure occurrence (for example, when the equipment failure probability reaches a certain specific threshold), when the remaining life of the equipment is lower than a certain value, or the acceleration interval of failure occurrence. Data points that meet the preset failure conditions are extracted from the equipment failure curve, and during the extraction process, the absolute value of the failure probability, the speed and trend of failure occurrence, etc. are considered, such as the sharp increase interval of the failure probability, to obtain initial failure prediction information. In addition, the actual usage environment and historical maintenance data of the equipment are further combined in all the initial failure prediction information extracted in the previous step for further refined screening to ensure that only data points closely related to the current working conditions and historical performance of the equipment are extracted as the equipment failure prediction information.

[0064] In the above-mentioned equipment failure prediction method based on the digital twin model, by inputting the equipment operation state data into the fixed failure calculation model and the usage failure calculation model respectively, the object fixed failure data and the object usage failure data of the target device are generated. Then, the temperature data of the equipment is used to correct the fixed failure data, and the load cycle data is used to correct the usage failure data, so as to more accurately reflect the failure trend of the equipment under different working environments. Further, the failure curve of the equipment is calculated using the two corrected failure data, and the failure prediction information that meets the preset failure conditions is extracted from it to realize the early warning of equipment failure. Through the above failure prediction mechanism, the real-time performance and accuracy in the prediction of equipment failure can be effectively improved, the reliability of equipment operation can be further significantly improved, the maintenance and repair cycle can be optimized, the downtime caused by sudden failures can be reduced, the maintenance cost can be lowered, and the overall operation and maintenance efficiency of the equipment can be improved, thus providing a strong guarantee for the long-term stable operation of the equipment.

[0065] In an exemplary embodiment, as Figure 3 shown, correcting the object fixed failure data according to the equipment operation temperature data to obtain the temperature-corrected fixed failure data, including steps 301 to 302. Among them:

[0066] Step 301, according to the equipment operation temperature data and the object fixed failure data, determine the life simulation environment parameters and life simulation duration parameters corresponding to the accelerated life simulation.

[0067] Among them, the life simulation environment parameters can be various variables used to simulate the actual working environment conditions of the equipment during the accelerated life test. These parameters include external factors such as temperature, humidity, vibration, and load, which will directly affect the aging speed and failure occurrence mode of the equipment. Especially in terms of temperature, the working conditions of the equipment at high temperatures usually lead to accelerated material aging, so temperature is the key environmental parameter in life simulation.

[0068] Among them, the life simulation duration parameter can be the time length used to simulate the operation of the device under specific environmental conditions in the accelerated life test. Different from the actual usage duration, the life simulation duration is usually calculated based on the acceleration factor, aiming to quickly obtain the trend of device aging and failure by shortening the test time. For example, in a high-temperature environment, the aging process of the device will be accelerated, and the simulation duration will be set according to this acceleration ratio.

[0069] Specifically, since the life simulation environment parameters usually involve the temperature stress borne by the device, considering the influence of high temperature on the device materials (for example, the acceleration effect of temperature on metal fatigue, plastic aging, etc.), and mapping these temperature data to an acceleration factor for simulating the working conditions in a high-temperature environment. Specifically, when the temperature is relatively high, the failure rate of the device will accelerate. Therefore, the model will be adjusted according to the temperature difference to set an acceleration factor, which is usually calculated through temperature-related models such as the Arrhenius equation, reflecting the influence of temperature on the device aging process. Therefore, according to the device operating temperature data and the object fixed failure data, the specific interval of the life simulation environment parameters corresponding to the accelerated life simulation is determined. At the same time, it is set based on the aging expectation of the device at a specific temperature. Usually, an accelerated usage time will be set to simulate the "accelerated usage cycle" that the device has experienced in this high-temperature environment as the life simulation duration parameter, ensuring that the simulated duration can reflect the failure trend of the device under actual high-temperature conditions.

[0070] Step 302: Perform an accelerated life simulation on the target device according to the life simulation environment parameters and the life simulation duration parameter to obtain temperature-corrected fixed failure data.

[0071] Among them, the accelerated life simulation can be a test method to predict the long-term service life of a device by artificially accelerating the device aging process. By artificially increasing environmental factors such as temperature and load in the laboratory environment, it simulates the usage situation of the device under extreme conditions, thereby accelerating the device aging process and revealing in advance the possible failure modes and failure times of the device.

[0072] Specifically, during the accelerated life simulation, by placing the device in a high-temperature environment or simulating high-temperature conditions (e.g., by heating the device or controlling the ambient temperature), the aging process that the device would experience over a longer period is simulated within a shorter time. At this time, factors such as the device's operating state, temperature, and load are dynamically monitored, and the operating conditions are set to be more severe than the normal working environment to accelerate the aging process of the device. By conducting accelerated tests on the device, the failure modes that might occur during long-term use can be completed within a shorter time, and then indicators such as the failure probability and failure rate of the device under these temperature conditions can be calculated. All the data will generate a temperature-corrected fixed failure dataset, which represents the impact of temperature changes on the occurrence of failures during the actual operation of the device. This usually includes continuously monitoring various key indicators during the operation of the device, such as temperature, pressure, vibration, etc., recording and analyzing the occurrence time and frequency of device failures. Finally, the temperature-corrected fixed failure data obtained through simulation can provide a more accurate basis for predicting future device failures, truly reflecting the accelerating effect of temperature on the device's life, and providing more detailed and accurate inputs for subsequent failure prediction models.

[0073] The calculation expression for realizing temperature-corrected fixed failure data in accelerated life simulation or actual production work is as follows:

[0074]

[0075]

[0076]

[0077] Wherein, is the object fixed failure data, is the initial inherent failure data, obtained from the device's user manual, where the data in the user manual is the result of the device manufacturer's tests on the device. T(t) is the device operating temperature data of the object device at time t , is the influence coefficient of temperature on the inherent failure data, determined by studying the impact of temperature on the device through temperature-accelerated aging tests in the laboratory, is the degradation coefficient related to time of the object device, determined by simulating extreme operating conditions (such as high temperature, high load, etc.) on the device through accelerated life tests, is the activation energy, is the Boltzmann constant, T is the temperature, is the load or stress, H is the humidity, is the sensitivity coefficient of load on the inherent failure data, determined by laboratory material fatigue tests (such as uniaxial tensile tests, cyclic loading tests) to determine the change in the failure rate of different materials under different loads, is the sensitivity coefficient of humidity to the inherent failure data, determined by studying the impact of humidity on the equipment through humidity accelerated aging tests in the laboratory. is the load effect. is the humidity effect. is the temperature effect. is the Arrhenius model failure data. is the temperature-corrected fixed failure data.

[0078] In this embodiment, by determining the environmental parameters and duration parameters for accelerated life simulation based on the equipment operating temperature data and fixed failure data, the failure process of the equipment under extreme conditions such as high temperature can be simulated in a relatively short time. Among them, the accelerated life simulation effectively reduces the time and cost required in actual testing, and at the same time can more quickly evaluate the impact of temperature on the fixed failure of the equipment. Further, by obtaining the temperature-corrected fixed failure data, it helps to predict the long-term stability and reliability of the equipment in the actual operating environment, thereby discovering potential failure risks in advance, optimizing the maintenance plan, extending the service life of the equipment, and improving the operating safety and economy of the equipment.

[0079] In an exemplary embodiment, as Figure 4 shown, the usage failure data of the object is corrected according to the equipment load cycle data to obtain the load-corrected usage failure data, including steps 401 to 402. Among them:

[0080] Step 401, according to the equipment load cycle data and the usage failure data of the object, determine the usage simulation environmental parameters and usage simulation mechanical parameters corresponding to the usage fatigue simulation.

[0081] Among them, the usage fatigue simulation can be a computer simulation method based on the actual workload and environmental conditions, used to predict the fatigue damage caused by repeated loads during the long-term use of the equipment. This simulation exposes the equipment to load cycles, simulates the stress, strain response of the equipment material, and the cumulative process of fatigue cracks. By repeatedly loading, the damage evolution of the equipment under various load conditions is simulated, and its fatigue life is finally predicted.

[0082] Among them, the usage simulation environmental parameters can be the working environmental conditions set for the equipment during the usage fatigue simulation, including factors such as load amplitude, load frequency, working temperature, humidity, etc. These environmental parameters reflect the working state and external conditions experienced by the equipment during actual use, and they have an important impact on the fatigue life and failure mode of the equipment.

[0083] Among them, the simulated mechanical parameters can be the parameters related to the mechanical properties of materials involved in fatigue simulation, such as the yield strength, fatigue limit, elastic modulus, stress-strain relationship, etc. of the material. These parameters reflect the mechanical behavior of the equipment material under different stress conditions and determine the cumulative rate of its fatigue damage and failure mode. By inputting these mechanical parameters into the fatigue simulation, the stress distribution of the equipment under load, the fatigue crack propagation of the material, and the final failure point can be accurately predicted.

[0084] Specifically, based on the equipment load cycle data and the object usage failure data of the target equipment, analyze information such as the load amplitude, load cycle frequency, and duration experienced by the equipment during actual operation, determine the fluctuation mode of the load and its impact on the fatigue damage of the equipment. Among them, a larger load amplitude and a higher frequency usually accelerate the fatigue damage of the equipment, so these factors will directly determine the loading conditions in the simulation. At the same time, by combining the object usage failure data, understand the actual failure prediction situation of the equipment under different load conditions without correction, so as to calculate the initial simulation parameters and make the simulation more conform to the usage state of the equipment in the real working condition. In addition, the influence of external environmental factors such as temperature and humidity on the fatigue characteristics of the equipment needs to be considered, and the initial simulation parameters are fine-tuned to obtain the simulated environment parameters for use.

[0085] At the same time, extract key working condition information from the equipment load cycle data, such as the load amplitude, load cycle frequency, load change mode (such as sine wave, impact loading, etc.) experienced by the equipment, and the distribution of these loads in the operation cycle. These working condition information reflect the load fluctuation characteristics borne by the equipment during actual use and are the core inputs for evaluating fatigue damage. Further, based on the working environment and material properties of the equipment, determine the simulated mechanical parameters for use, including the yield strength, fatigue limit, elastic modulus, stress-strain curve, etc. of the material. These mechanical parameters are crucial for predicting the fatigue behavior of the material under repeated loading. For example, the fatigue life of the material usually shortens with the increase of the load frequency and amplitude. By integrating the load cycle data, the simulated environment parameters for use, and the simulated mechanical parameters for use, a fatigue simulation environment is constructed, which reflects the behavior of the equipment under the actual working load. During the process of using fatigue simulation, fatigue analysis methods such as the rain flow counting method, S-N curve, or FEM (finite element analysis) model are usually used.

[0086] Step 402, perform fatigue simulation for use on the target equipment according to the simulated environment parameters for use and the simulated mechanical parameters for use, and obtain the load-corrected usage failure data.

[0087] Specifically, based on the use of simulation environment parameters and simulation mechanical parameters, the actual fatigue simulation process is entered. At this time, through computer simulation software or methods such as the rain flow counting method, S-N curve, or FEM (finite element analysis), the equipment load cycle data of the equipment is combined with the material mechanics parameters to simulate the fatigue damage of the equipment under repeated loads. The simulation process will simulate multiple cycles of the load, calculate the stress and strain effects of each load on the internal materials of the equipment, and gradually accumulate damage. For example, under high loads, microcracks may appear in the equipment components, and with more load cycles, these cracks will gradually expand and lead to equipment failure. Through simulation, the degree of fatigue damage experienced by the equipment during long-term use can be quantified, and the fatigue life of each component and the time of failure occurrence can be obtained. The simulation results will generate load-corrected usage failure data, which accurately reflects the potential failure risk caused by fatigue damage of materials due to load cycles under the actual working conditions of the equipment. The load-corrected usage failure data is more in line with the actual usage environment of the equipment compared to the uncorrected data.

[0088] For the calculation expression of load-corrected usage failure data in fatigue simulation or actual production work:

[0089]

[0090]

[0091]

[0092] Among them, is the object usage failure data, is the initial usage failure data, obtained by statistically analyzing the usage of other equipment with the same usage environment and usage method. T(t) is the equipment operating temperature data of the object equipment at time t , is the influence coefficient of temperature on usage failure data, determined by studying the influence of temperature on the equipment through temperature accelerated aging tests in the laboratory, is the degradation coefficient related to time of the object equipment, determined by simulating extreme working conditions (such as high temperature, high load, etc.) of the equipment through accelerated life tests. T is the temperature, is the load or stress, H is the humidity, is the sensitivity coefficient of humidity on usage failure data, determined by studying the influence of humidity on the equipment through humidity accelerated aging tests in the laboratory, is the equipment load cycle data, is the growth of failure data of the object equipment degraded over time, is the time degradation rate, obtained by fitting the usage of other equipment with the same usage environment and usage method. is the correction factor related to the device load cycle data, is the material constant obtained by conducting fatigue strain tests on the material in the laboratory (such as tension-compression cycle tests), is the influence coefficient of temperature on the strain of the device material, determined through fatigue tests under temperature changes in the laboratory, is the influence coefficient of load on the strain of the device material, determined by the influence of the load amplitude on the strain in fatigue experiments under different loads, is the influence coefficient of humidity on the strain of the device material, determined by the influence on the device in humidity accelerated tests. c is the Coffin–Manson constant obtained through experimental tests, where the experiments usually include conducting low-cycle fatigue tests on the material and recording the fatigue life of the material at different strain amplitudes, is the total strain amplitude, is the load correction using failure data.

[0093] In this embodiment, by determining the environmental parameters and mechanical parameters for fatigue simulation based on the device load cycle data and the usage failure data, the fatigue behavior of the device under long-term load can be accurately simulated under conditions closer to the actual working conditions. Such simulation can reveal the specific impacts of load fluctuations, frequent undulations, and the physical responses of the device on the occurrence of failures, thereby better understanding the fatigue damage accumulation of the device during actual use. The load correction using failure data obtained through this process can effectively supplement the deficiencies of traditional failure prediction models, provide more accurate failure prediction results, help identify potential failures caused by load cycles in advance, optimize the device maintenance cycle, reduce sudden failures and downtime, and improve the reliability and service life of the device.

[0094] In an exemplary embodiment, as Figure 5 shown, based on the temperature correction of the fixed failure data and the load correction of the usage failure data, the failure data of the target device on the usage time axis is calculated to obtain the device failure curve, including steps 501 to 503. Among them:

[0095] Step 501: Using each usage duration data point on the usage time axis as each integral upper limit value, the temperature correction of the fixed failure data and the load correction of the usage failure data are respectively integrated to obtain each integral fixed failure data and each integral usage failure data.

[0096] Among them, the usage duration data points can be each time node in the entire process from the device being put into use to the occurrence of a failure or the end of operation, located on the usage time axis. Each data point represents the operating state of the device at a certain specific duration, such as hours, days, or months, etc.

[0097] Among them, the integrated fixed-fault data can be the result of cumulative calculation of the fixed-fault data corrected by temperature based on the usage duration data points of the device. At each duration data point, integration is used to reflect the cumulative risk of the device due to fixed faults during this time period. The integration process reflects the accumulation of the device in terms of the fixed-fault probability over time.

[0098] Among them, the integrated usage-fault data can be the result of cumulative calculation of the usage-fault data corrected by load based on the usage duration data points of the device. Different from the fixed-fault data, the usage faults are usually related to the load cycles that the device undergoes during operation. The process of integrating the usage-fault data reflects the gradually accumulated fatigue damage and fault risk probability of the device under different load cycles.

[0099] Specifically, the life simulation duration parameter is used as a reference duration to define the usage time axis of the device. The usage time axis covers the entire time period from the start of the device's operation to the occurrence of a fault. To ensure the sufficiency of the usage time axis, the time length of the usage time axis is greater than the life simulation duration. Each data point on the usage time axis represents the operation duration of the device at a certain moment, such as hours, days, months, etc. These duration data points will be used as the upper limits of integration to calculate the cumulative situation of the device's faults at different time points. The specific operation is to associate the fixed-fault data corrected by temperature and the usage-fault data corrected by load with each duration point on the time axis respectively. The integration process is to sum up the fixed-fault data corrected by temperature and the usage-fault data corrected by load of the device in each time period, indicating the cumulative fault impact on the device during this time. Through integration, the fault risk probability and cumulative damage value of the device at different usage durations can be obtained, and the integrated fixed-fault data and the integrated usage-fault data can be obtained.

[0100] Step 502: Fit the integrated fixed-fault data and the integrated usage-fault data respectively to obtain a fixed-fault fitting curve and a usage-fault fitting curve.

[0101] Among them, the fixed-fault fitting curve can be a curve obtained by mathematically fitting the integrated fixed-fault data, which is used to describe the trend of damage accumulation due to fixed faults of the device under specific operating conditions.

[0102] Among them, the usage-fault fitting curve can be a curve obtained by mathematically fitting the integrated usage-fault data, indicating the trend of fault accumulation due to fatigue damage of the device during the usage load cycle.

[0103] Specifically, since it is necessary to fit the temperature-corrected fixed fault data and the load-corrected used fault data separately, an appropriate mathematical fitting model (such as an exponential model, a power-law model, or a polynomial fit) is selected to represent the relationship between the data and time. These models will use the least squares method or other optimization algorithms to find the best-fit curve, so that the fit curve can conform to the actual fault data as much as possible. The obtained fixed fault fit curve and the used fault fit curve respectively describe the fault evolution process of the device under the influence of temperature and load.

[0104] Step 503: Superimpose the fixed fault fit curve and the used fault fit curve to obtain the device failure curve.

[0105] Specifically, the key to superimposing the fixed fault fit curve and the used fault fit curve is to calculate how the temperature and load effects of the device failure act together and accumulate at each time node. In order to more accurately reflect the combined influence of the temperature and load correction factors, a non-linear superposition mechanism is adopted, which takes into account the complex interaction relationships between different types of faults (such as fixed faults and used faults). In this process, first, the fault data of each fit curve needs to be normalized to eliminate the dimensional differences and influence strength differences between different fault types. Then, by establishing a weight coefficient or correction factor, according to the working state, load conditions, and temperature changes of the device, the summation method at each time point is dynamically adjusted; especially in the case of the interaction between load and temperature, this summation is not just a simple mathematical summation, but also requires the introduction of multiplication factors or exponential superposition, and these factors can be adjusted according to the device's historical fault data and actual working conditions, so that under extreme load or temperature conditions, the increase in fault risk shows a more significant non-linear growth trend, and the device failure curve is obtained.

[0106] In this embodiment, by taking each duration data point on the usage time axis as the upper limit of integration and integrating the temperature-corrected fixed fault data and the load-corrected used fault data, the fault accumulation situation at different usage durations can be obtained, which can more accurately reflect the fault development trend of the device during long-term use. Further, by fitting the integrated data, the fit curves of fixed faults and used faults can be obtained, which can provide a more accurate prediction model for device failure. Superimposing these two fit curves, the finally obtained device failure curve can comprehensively consider the influence of temperature and load cycles on the device life, thus providing a comprehensive reference for fault prediction. It can identify in advance the possible faults of the device in the future usage cycle, optimize the maintenance and replacement plan, reduce the device downtime risk, and improve the operation reliability and safety of the device.

[0107] In an exemplary embodiment, such as Figure 6As shown, the fixed failure fitting curve and the usage failure fitting curve are superimposed to obtain the equipment failure curve, including steps 601 to 602. Among them:

[0108] Step 601, perform a correlation analysis on the fixed failure fitting curve and the usage failure fitting curve to determine the curve correlation data.

[0109] Among them, the correlation analysis can be used to explore the relationship and the degree of mutual influence between two or more variables through statistical methods. In equipment failure prediction, the correlation analysis is used to study the relationship between the fixed failure data corrected by temperature and the usage failure data corrected by load, aiming to identify their combined effects in the equipment failure process.

[0110] Among them, the curve correlation data can be the results obtained from the correlation analysis, which reflects the mutual relationship and the degree of influence between the fixed failure fitting curve and the usage failure fitting curve. These data include the correlation coefficient, interaction strength, and the degree of influence of one curve on the other curve at different time points between the two curves, etc.

[0111] Specifically, it is necessary to analyze the mutual relationship between the fixed failure fitting curve and the usage failure fitting curve. The purpose of the mutual relationship analysis is to clarify how the two curves act together in the equipment failure process. Specific analysis methods include, for example, performing time synchronization processing on the curves, that is, aligning the corresponding time points on the two curves to ensure their performance in the same time dimension; then using statistical methods such as correlation analysis or covariance analysis to quantify the correlation strength between the two curves. If the change trends of the two curves are highly consistent in some periods, it may mean that during these periods, the contributions of the fixed failure and the usage failure to the equipment failure are mutually reinforcing. In addition, methods such as multiple regression analysis or principal component analysis can also be used to explore the key factors affecting the correlation between the two curves, such as external conditions such as the load change and environmental temperature of the equipment. Through the correlation analysis, the curve correlation data can be obtained, including the intersection points, correlation coefficients, and their combined effects on the equipment failure at different stages of the two curves.

[0112] Step 602, according to the curve correlation data, perform a non - linear superposition on the fixed failure fitting curve and the usage failure fitting curve to obtain the equipment failure curve.

[0113] Among them, the non - linear superposition can be to perform addition or fusion in a non - linear manner when combining two or more influencing factors to reflect their complex interaction relationship.

[0114] Specifically, based on the curve correlation data obtained from the correlation analysis, the fixed-fault fitting curve and the usage-fault fitting curve are nonlinearly superimposed. The key to the nonlinear superposition lies in the fact that the influence of the two curves on equipment failure is not a simple additive relationship, but there is a certain complex interaction. In the specific implementation process, it is necessary to use the nonlinear superposition algorithm to dynamically adjust the weights of the two curves according to the curve correlation data. For example, the curve correlation data shows that within certain specific time periods, temperature has a greater impact on fixed faults, while load has a stronger impact on usage faults. Then, within these time periods, the formula for nonlinear superposition should give more weight to the fixed-fault fitting curve, and at the same time adjust the superposition ratio in other time periods. Among them, in the process of the nonlinear superposition algorithm, methods such as product superposition, exponential superposition, or logarithmic superposition can be adopted. For example, by constructing a product model, the total failure probability of the equipment is jointly determined by the influences of fixed faults and usage faults, and the contribution degree of each curve changes dynamically according to the actual working conditions. Through the nonlinear superposition algorithm, the nonlinear characteristics between the fixed-fault fitting curve and the usage-fault fitting curve are mutually enhanced or mutually suppressed to obtain the equipment failure curve.

[0115] In this embodiment, by performing correlation analysis on the fixed-fault fitting curve and the usage-fault fitting curve, the mutual influence of temperature correction and load correction on the development of equipment faults can be revealed. The determination of curve correlation data can help judge the relationship strength and interaction between the two, providing a basis for subsequent nonlinear superposition. Based on these correlation data, the two curves are nonlinearly superimposed, and the obtained equipment failure curve can more accurately reflect the fault evolution process of the equipment under complex working conditions. It not only improves the accuracy of fault prediction but also comprehensively considers the superposition effect of multiple influencing factors, providing more reliable data support for the maintenance decision-making of the equipment, thereby optimizing the maintenance strategy, reducing the equipment failure rate, extending the service life of the equipment, and ensuring the stability and safety of the production process.

[0116] In an exemplary embodiment, as Figure 7 shown, according to the curve correlation data, the fixed-fault fitting curve and the usage-fault fitting curve are nonlinearly superimposed to obtain the equipment failure curve, including steps 701 to 706. Among them:

[0117] Step 701, determine the exponential superposition parameter, logarithmic superposition parameter, and power-law superposition parameter according to the fixed-fault fitting curve and the usage-fault fitting curve.

[0118] Among them, the exponential superposition parameters can be coefficients used to adjust the curve shape when performing exponential model superposition. These parameters control the fault growth rate and acceleration degree, and usually include the reference fault value and the growth factor (i.e., the exponent value of the exponent), which determine how factors such as temperature and load affect the acceleration trend of faults during the operation of the device.

[0119] Among them, the logarithmic superposition parameters can be coefficients used to adjust the curve shape when performing logarithmic model superposition. Usually include the reference fault value and the logarithmic growth coefficient, and these parameters define the cumulative characteristics of the device fault under changes in time or working conditions.

[0120] Among them, the power-law superposition parameters can be parameters used to adjust the curve shape when performing power-law model superposition. These parameters include the power exponent and the constant term, which control the degree of non-linear change during the device fault growth process.

[0121] Specifically, perform curve feature extraction on the fitted fixed-fault fitting curve and the used-fault fitting curve to analyze their growth rates, change forms, and potential interactions. For example, if the growth of the device fault shows exponential acceleration, logarithmic accumulation, or power-law non-linear growth, the system will automatically determine the exponential superposition parameters, logarithmic superposition parameters, and power-law superposition parameters corresponding to the superposition algorithm suitable for each curve through optimization algorithms (such as the least squares method, maximum likelihood estimation, etc.).

[0122] Step 702, respectively use the exponential superposition parameters, logarithmic superposition parameters, and power-law superposition parameters to perform non-linear superposition on the fixed-fault fitting curve and the used-fault fitting curve to obtain the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve.

[0123] Among them, the exponential superposition curve can be a curve generated by combining different fault data (such as fixed-fault data and used-fault data) through an exponential model. It is mainly used to describe the acceleration process of device faults and is usually applicable to rapid fault growth caused by factors such as high temperature and high load.

[0124] Among them, the logarithmic superposition curve can be a curve generated by combining the fixed-fault data and used-fault data of the device through a logarithmic model. This curve is used to describe the faults gradually accumulated during the long-term operation of the device.

[0125] Among them, the power-law superposition curve can be a curve obtained by combining the fixed-fault data and used-fault data through a power-law model, mainly used to describe the growth mode of device faults under specific non-linear working conditions. The power-law superposition curve can capture complex fault modes caused by load fluctuations and extreme environmental changes.

[0126] Specifically, the exponential superposition parameter, logarithmic superposition parameter, and power-law superposition parameter are respectively applied to the exponential superposition algorithm, logarithmic superposition algorithm, and power-law superposition algorithm to perform non-linear superposition on the fixed-fault fitting curve and the used-fault fitting curve. During the non-linear superposition process, the exponential superposition algorithm will enhance the accelerating growth trend of the fault and is applicable to rapid failures caused by temperature changes; the logarithmic superposition algorithm is suitable for simulating faults gradually accumulated during long-term stable operation; the power-law superposition algorithm can handle complex non-linear fault growth and is usually applicable to the non-uniform effects of load fluctuations on equipment. During actual calculation, the system combines the numerical values of the two curves according to the formula of each superposition model to obtain three different superposition curves, namely the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve.

[0127] Step 703: Calculate the errors of the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve respectively to obtain the curve error data of each curve.

[0128] Among them, the curve error data can be the data obtained by comparing the difference between the superposition curve and the actual fault data. Error calculation usually adopts statistical methods such as mean square error (MSE) or mean absolute error (MAE). The error data reflects the accuracy of different superposition models (such as exponential, logarithmic, and power-law superposition curves) in fitting the actual fault data.

[0129] Specifically, for the errors between the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve pairwise, error measurement standard algorithms such as mean square error (MSE), mean absolute error (MAE), or relative error (RE) and other indicators are usually used to quantify the difference between each superposition curve and the actual fault data, where the actual fault data comes from equipment under the same environmental conditions and the same usage conditions, and the usage time of this equipment is much longer than that of the target equipment. The system compares each superposition curve with the actual fault data, calculates the curve error data of each superposition curve, and then understands the accuracy of different superposition methods for fault prediction; the smaller the error, the more consistent the superposition model is with the law of actual equipment faults.

[0130] Step 704: When the differences between the curve error data are all less than the preset error standard, perform complexity analysis or generalization ability analysis on the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve respectively to obtain the complexity analysis data or generalization analysis data of each curve.

[0131] Among them, the preset error standard can be the tolerance range used to measure the fitting effect of each superimposed curve when selecting a model. Usually during the model fitting process, an error range is set, requiring that the error difference between the superimposed curve and the actual fault data does not exceed this standard. If the error difference exceeds the preset error standard, it indicates that the fitting effect of this model is poor.

[0132] Among them, complexity analysis can be the process of evaluating the complexity of the model when predicting faults. During this process, the main focus is on the number of parameters of the model, the computational amount, and the adaptability to data. A model with a higher complexity may be too complex and prone to overfitting; while a model with too low complexity may not be able to capture the key features in the data.

[0133] Among them, generalization ability analysis can be the evaluation of the prediction performance of the model on unknown data or new working conditions. It examines whether the model can adapt to different equipment operating environments and working conditions, so as to effectively predict future faults. Through techniques such as cross-validation, generalization ability analysis helps to judge the stability and robustness of the model.

[0134] Among them, complexity analysis data can be the data obtained by evaluating the complexity of the model, usually including the number of parameters of the model, the computational complexity, and the computational efficiency of the model.

[0135] Among them, generalization analysis data can be the data obtained by generalization ability analysis, usually including the prediction accuracy, stability of the model on unseen data, and the adaptability to different working conditions.

[0136] Specifically, after calculating the errors of the exponential superimposed curve, logarithmic superimposed curve, and power-law superimposed curve respectively, if the error differences of the exponential superimposed curve, logarithmic superimposed curve, and power-law superimposed curve are all less than the preset error standard (for example, the error differences are within the tolerance range), it indicates that these three curves can all fit the actual data well, and further complexity analysis or generalization ability analysis needs to be carried out. Among them, complexity analysis mainly examines the complexity of each model, such as the number of parameters of the model, the computational amount, etc., to avoid overfitting of the model and ensure its strong structural adaptability and simplicity. Generalization ability analysis examines the performance of the model on unseen data, that is, whether the model can effectively predict the failure of the equipment under different working conditions. The system will carry out generalization ability analysis through techniques such as cross-validation and leave-one-out method, evaluate the adaptability and robustness of each superimposed curve under different working conditions, and obtain the complexity analysis data or generalization analysis data of each superimposed curve.

[0137] Step 705, select the superimposed curve with the minimum complexity analysis data from the exponential superimposed curve, logarithmic superimposed curve, and power-law superimposed curve as the equipment failure curve.

[0138] Specifically, compare the complexity analysis data of the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve, and select the curve with the minimum complexity among the various complexity analysis data as the final equipment failure curve. For example, in large-scale equipment monitoring, selecting the model with the lowest computational complexity can reduce the computational resource occupancy of the system and improve the response speed while ensuring the prediction accuracy.

[0139] Step 706: Select the superposition curve with the maximum generalization analysis data from the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve as the equipment failure curve.

[0140] Specifically, compare the generalization ability scores of the exponential superposition curve, logarithmic superposition curve, and power-law superposition curve, and select the superposition curve with the highest score, that is, the curve with the maximum generalization ability data, as the final equipment failure curve, ensuring that the selected curve can still maintain a high prediction accuracy and adaptability under different operating environments or working conditions, especially when facing different types of equipment or external changes.

[0141] In this embodiment, by determining the exponential, logarithmic, and power-law superposition parameters for the fixed-fault fitting curve and the used-fault fitting curve and performing non-linear superposition, three different types of failure curves can be obtained. These curves can reflect the failure trends of the equipment under complex working conditions from different perspectives. By calculating the errors and analyzing the complexity of these curves, the fitting accuracy and computational complexity of each curve can be effectively evaluated, ensuring that the selected failure curve achieves the best balance between accuracy and computational resource usage. In addition, the generalization ability analysis ensures that the selected curve has good adaptability and stability under different usage conditions. Finally, by selecting the superposition curve with the minimum complexity or the maximum generalization ability, it can be ensured that the obtained equipment failure curve is not only accurate in theory but also has strong robustness and prediction accuracy in practical applications. This method greatly improves the reliability of equipment fault prediction, helps to identify potential risks in advance, optimize the equipment maintenance strategy, reduce the probability of faults, and thus improve the usage efficiency and operating safety of the equipment.

[0142] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0143] Based on the same inventive concept, an embodiment of the present application also provides a device for predicting equipment faults based on a digital twin model for implementing the method for predicting equipment faults based on a digital twin model involved above. The implementation solutions provided by this device to solve problems are similar to the implementation solutions described in the above method. Therefore, the specific limitations in one or more embodiments of the device for predicting equipment faults based on a digital twin model provided below can refer to the limitations on the method for predicting equipment faults based on a digital twin model in the above text, and will not be repeated here.

[0144] In an exemplary embodiment, as Figure 8 shown, a device for predicting equipment faults based on a digital twin model is provided, including: an equipment data acquisition module 801, a fault data calculation module 802, a fault data correction module 803, a failure curve generation module 804, and an equipment fault prediction module 805. Each module in the above device for predicting equipment faults based on a digital twin model can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules to implement any step of the method for predicting equipment faults based on a digital twin model.

[0145] In an exemplary embodiment, a computer device is provided. This computer device can be a server, and its internal structure diagram can be as Figure 9 shown. This computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. When the computer program is executed by the processor, it is used to implement a method for predicting equipment faults based on a digital twin model.

[0146] Those skilled in the art can understand, Figure 9The structure shown is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0147] In one embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0148] In one embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0149] In one embodiment, a computer program product or a computer program is provided. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steps in the above method embodiments.

[0150] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0151] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.

[0152] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0153] The above-described embodiments merely represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A device fault prediction method based on a digital twin model, characterized in that The method includes: When the device monitoring twin model fails to detect fault information of the target device, obtaining the device operation status data, device operation temperature data, and device load cycle data of the target device; Substituting the device operation status data into the fixed fault calculation model and the usage fault calculation model of the target device respectively to obtain target fixed fault data and target usage fault data; Determining the life simulation environment parameters and life simulation duration parameters corresponding to the accelerated life simulation according to the device operation temperature data and the target fixed fault data; Performing the accelerated life simulation on the target device according to the life simulation environment parameters and the life simulation duration parameters to obtain the temperature-corrected fixed fault data; Determining the usage simulation environment parameters and usage simulation mechanical parameters corresponding to the usage fatigue simulation according to the device load cycle data and the target usage fault data; Performing the usage fatigue simulation on the target device according to the usage simulation environment parameters and the usage simulation mechanical parameters to obtain the load-corrected usage fault data; Calculating the failure data of the target device on the usage time axis according to the temperature-corrected fixed fault data and the load-corrected usage fault data to obtain a device failure curve; Extracting the data points that meet the preset failure conditions from the device failure curve as the device fault prediction information of the target device.

2. The method according to claim 1, wherein The calculation expression of the temperature-corrected fixed fault data is: Among them, is the object fixed failure data, is the initial inherent failure data, T(t) is the equipment operating temperature data of the object equipment at time t ; is the influence coefficient of temperature on the inherent failure data, is the degradation coefficient related to time of the object equipment, is the activation energy, is the Boltzmann constant, T is the temperature, is the load or stress, H is the humidity, is the sensitivity coefficient of load to the inherent failure data, is the sensitivity coefficient of humidity to the inherent failure data, is the load effect, is the humidity effect, is the temperature effect, is the Arrhenius model failure data, is the temperature-corrected fixed failure data. It should be noted that in the original text, there is a semicolon missing in line 6. I added it in the translation for better logical integrity.

3. The method according to claim 1, wherein The calculation expression of the load-corrected usage fault data is: Among them, Use failure data for the object, is the initial use failure data, T(t) is the device operating temperature data of the object device at time t of, is the influence coefficient of temperature on use failure data, is the degradation coefficient related to the object device and time, T is the temperature, is the load or stress, H is the humidity, is the sensitivity coefficient of humidity to use failure data, is the device load cycle data, is the growth of failure data due to the degradation of the object device over time, is the time degradation rate, is the correction coefficient related to the device load cycle data, is the material constant, is the influence coefficient of temperature on the strain of the device material, is the influence coefficient of load on the strain of the device material, is the influence coefficient of humidity on the strain of the device material, c is the Coffin–Manson constant, is the total strain amplitude, is the load-corrected use failure data.

4. The method according to claim 1, wherein The calculating the failure data of the target device on the usage time axis according to the temperature-corrected fixed fault data and the load-corrected usage fault data to obtain a device failure curve includes: Taking each usage duration data point on the usage time axis as each integral upper limit value, and integrating the temperature-corrected fixed fault data and the load-corrected usage fault data respectively to obtain each integral fixed fault data and each integral usage fault data; Fitting each of the integral fixed fault data and each of the integral usage fault data respectively to obtain a fixed fault fitting curve and a usage fault fitting curve; Superimposing the fixed fault fitting curve and the usage fault fitting curve to obtain the device failure curve.

5. The method according to claim 4, characterized in that, The superimposing the fixed fault fitting curve and the usage fault fitting curve to obtain the device failure curve includes: Performing a correlation analysis on the fixed fault fitting curve and the usage fault fitting curve to determine curve correlation data; Nonlinearly superimposing the fixed fault fitting curve and the usage fault fitting curve according to the curve correlation data to obtain the device failure curve.

6. The method according to claim 5, characterized in that The nonlinearly superimposing the fixed fault fitting curve and the usage fault fitting curve according to the curve correlation data to obtain the device failure curve includes: Determining an exponential superimposing parameter, a logarithmic superimposing parameter, and a power-law superimposing parameter according to the fixed fault fitting curve and the usage fault fitting curve; Using the exponential superposition parameter, the logarithmic superposition parameter, and the power-law superposition parameter respectively, perform non-linear superposition on the fixed-fault fitting curve and the used-fault fitting curve to obtain an exponential superposition curve, a logarithmic superposition curve, and a power-law superposition curve; Calculate the error of the exponential superposition curve, the logarithmic superposition curve, and the power-law superposition curve respectively to obtain the error data of each curve; When the differences between the error data of each curve are all less than the preset error standard, perform complexity analysis and generalization ability analysis on the exponential superposition curve, the logarithmic superposition curve, and the power-law superposition curve respectively to obtain the complexity analysis data and the generalization analysis data of each; Select the superposition curve with the minimum complexity analysis data among the exponential superposition curve, the logarithmic superposition curve, and the power-law superposition curve as the device failure curve; Alternatively, select the superposition curve with the maximum generalization analysis data among the exponential superposition curve, the logarithmic superposition curve, and the power-law superposition curve as the device failure curve.

7. An apparatus for predicting equipment failures based on a digital twin model, characterized in that, The device includes: A device data acquisition module, configured to acquire the device operation status data, the device operation temperature data, and the device load cycle data of the target device when the device monitoring twin model fails to detect the fault information of the target device; A fault data calculation module, configured to substitute the device operation status data into the fixed-fault calculation model and the used-fault calculation model of the target device respectively to obtain the target fixed-fault data and the target used-fault data; A fault data correction module, configured to determine the life simulation environment parameters and the life simulation duration parameters corresponding to the accelerated life simulation according to the device operation temperature data and the target fixed-fault data; Perform the accelerated life simulation on the target device according to the life simulation environment parameters and the life simulation duration parameters to obtain the temperature-corrected fixed-fault data; The fault data correction module is further configured to determine the use simulation environment parameters and the use simulation mechanical parameters corresponding to the use fatigue simulation according to the device load cycle data and the target used-fault data; Perform the use fatigue simulation on the target device according to the use simulation environment parameters and the use simulation mechanical parameters to obtain the load-corrected used-fault data; A failure curve generation module, configured to calculate the failure data of the target device on the use time axis according to the temperature-corrected fixed-fault data and the load-corrected used-fault data to obtain a device failure curve; A device fault prediction module, configured to extract the data points that meet the preset failure conditions from the device failure curve as the device fault prediction information of the target device.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

Images