Embedded computer predictive maintenance control system based on digital twinning

Abstract

The invention discloses an embedded computer predictive maintenance control system based on digital twinning, which relates to the technical field of computers, and comprises a data fusion modeling module for constructing a high-precision digital model by using a digital twinning body; the fault anomaly detection module is used for judging whether the junction temperature of the chip exceeds an early warning critical point or not, giving out early warning, extracting a vibration signal energy entropy value by adopting wavelet packet decomposition, judging whether the energy entropy value exceeds a preset value or not, and judging that the embedded computer is abnormal if the energy entropy value exceeds the preset value; the prediction equipment maintenance module is used for dynamically adjusting a maintenance period by adopting reinforcement learning; and the man-machine collaborative decision module obtains machine evaluation, obtains man-machine collaborative evaluation based on the high-precision digital model, and combines the machine evaluation and the man-machine collaborative evaluation to obtain man-machine collaborative evaluation. The method has the advantages that the problem of the embedded computer can be predictively predicted, the maintenance period is dynamically adjusted, and predictive maintenance of the embedded computer is achieved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an embedded computer predictive maintenance control system based on digital twins. Background Art

[0002] Industrial embedded computers are equipped with embedded operating systems, which are stable, non-crashing, and virus-proof. They have independent embedded system disks, do not require hard disks as system disks, have no virtual memory, are not afraid of power failures damaging the operating system, and can be used as soon as they are powered on. This type of computer requires a maintenance device when in use.

[0003] If the maintenance device is used for a short period of time, the maintenance cost will be too high. If the maintenance device is used for a long period of time, the possibility of embedded computer failure will increase. The maintenance of embedded computers is based on the premise that the embedded computer has no failure. The reason for maintenance is to reduce the possibility of failure. If the embedded computer fails, the maintenance will lose its meaning. Summary of the invention

[0004] In order to solve the above technical problems, an embedded computer predictive maintenance control system based on digital twins is provided. This technical solution solves the problem that the maintenance cycle is difficult to determine and the embedded computer fails before maintenance proposed in the above background technology.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: Embedded computer predictive maintenance control system based on digital twin, including: A data fusion modeling module, wherein the data fusion modeling module sets sensors to monitor the temperature and vibration data of the embedded computer in real time, uploads the data to the cloud, and uses the digital twin to build a high-precision digital model; A fault anomaly detection module, which evaluates the junction temperature change of the chip based on a high-precision digital model, sets a junction temperature warning critical point, and issues a warning if the junction temperature of the chip exceeds the warning critical point. Based on the high-precision digital model, the module uses wavelet packet decomposition to extract the energy entropy value of the vibration signal to determine whether the energy entropy value exceeds a preset value. If so, it is determined that an abnormality has occurred in the embedded computer; A predictive equipment maintenance module, which uses reinforcement learning to dynamically adjust the maintenance cycle based on the prediction of the remaining life of the equipment; A human-machine collaborative decision-making module deploys a lightweight CNN model on the device side to achieve local real-time analysis of the vibration spectrum, obtain a machine evaluation, obtain a human evaluation based on a high-precision digital model, merge the machine evaluation and the human evaluation to obtain a human-machine collaborative evaluation, and perform maintenance control.

[0006] Preferably, the step of setting a sensor to monitor the temperature and vibration data of the embedded computer in real time, uploading the data to the cloud, and using the digital twin to build a high-precision digital model includes: Obtain the external dimensions of the embedded computer and construct a three-dimensional model of the embedded computer; Obtain the hardware component model, usage time and software composition of the embedded computer, and build a preliminary high-precision digital model of the embedded computer based on the three-dimensional model of the embedded computer; Obtain the temperature, vibration, and electrical parameter data of the embedded computer when it is running, and record them as real parameters; Use the preliminary high-precision digital model to simulate the normal operation of the embedded computer, and record the temperature, vibration, and electrical parameter data of the high-precision digital model as simulation parameters; Obtaining the normal fluctuation range of each parameter when the embedded computer is operating normally, calculating the length of the normal fluctuation range of each parameter, and recording half of the length of the normal fluctuation range of each parameter as the fluctuation preset value of each parameter; Compare the real parameters with the simulated parameters, calculate the difference between the real parameters and the simulated parameters, list each parameter as the target parameter one by one, and determine whether the difference of the target parameter is greater than the preset value of the fluctuation of the target parameter. If so, record the target parameter as the influencing parameter; Adjust the hardware model and usage time of the high-precision digital model, as well as the software composition, and record the changes in various parameters of the high-precision digital model; The hardware model and usage time of the high-precision digital model are adjusted into numbers according to the release year of the model and the length of usage time, which are recorded as independent variable parameters, and the degree of change of each parameter is quantified into numbers, which are recorded as dependent variable parameters; Based on the changes in various parameters of the high-precision digital model, the Pearson correlation coefficient between the independent variable parameters and the dependent variable parameters is calculated to obtain the correlation matrix between the independent variable parameters and the dependent variable parameters; Compare the difference between the influencing parameter and the true parameter, and determine the adjustment degree of the independent variable parameter corresponding to the influencing parameter based on the product of the Pearson coefficient of the correlation matrix of the influencing parameter and the difference; Summarize the adjustment degree of the independent variable parameter corresponding to at least one influencing parameter to obtain adjustment details of hardware and software; Based on the required hardware and software adjustment details, the preliminary high-precision digital model is adjusted to obtain a high-precision digital model; The Pearson correlation coefficient is a statistic used to measure the degree of linear correlation between two continuous variables, and its value range is -1 to 1.

[0007] Preferably, the step of evaluating the change in junction temperature of the chip and setting a junction temperature warning critical point includes: Get the temperature and air flow of the environment where the embedded computer is located; The average temperature at the chip node in the high-precision digital model is obtained by using the data in the high-precision digital model, which is recorded as the junction temperature of the chip; Obtain the time and load of common software used by users on embedded computers; Get the junction temperature when the user uses all common software at the same time, and record it as the maximum junction temperature; Evaluate the chip aging rate and performance when the maximum junction temperature is exceeded, and plot the change in junction temperature with the chip aging rate and performance; Find the point with the smallest derivative on the image and record the corresponding junction temperature as the warning critical point.

[0008] Preferably, the extracting the energy entropy value of the vibration signal by wavelet packet decomposition and judging whether the energy entropy value exceeds a preset value, and if so, judging that an abnormality occurs in the embedded computer comprises: Obtain vibration signals of high-precision digital models; Decomposing the vibration signal by using wavelet packets to obtain at least one divided frequency segment; Determine whether the decomposition layer number of the vibration signal decomposed by the wavelet packet belongs to the standard decomposition layer number interval. If yes, calculate the energy entropy value of the vibration signal using the entropy value calculation formula. If no, obtain the vibration signal of the high-precision digital model again and decompose the vibration signal using the wavelet packet. Determine whether the energy entropy value exceeds a preset value, and if so, determine that an abnormality occurs in the embedded computer; The entropy value calculation formula is: , In the formula, H is the entropy energy, is the ratio of the number of frequency bands with the frequency of the ith frequency band to the number of all frequency bands, and n is the number of frequency bands after the vibration signal wavelet packet decomposition; The wavelet packet decomposition is to pass the signal through a series of filters with different center frequencies but the same bandwidth to obtain a signal of at least one frequency band; The steps for obtaining the preset value and the standard decomposition layer number interval are as follows: Obtain at least one vibration frequency during at least one time period when the embedded computer is operating normally; Decomposing at least one vibration frequency by using wavelet packets to obtain the decomposition layer number and energy entropy value of at least one vibration frequency; Draw a bar chart of the number of decomposition layers of vibration frequency and energy entropy value; Based on the bar chart, find out the decomposition layer interval corresponding to the concentrated distribution of energy entropy values, record it as the standard decomposition layer interval, and record the average entropy value corresponding to the concentrated distribution of energy entropy values ​​as the preset value.

[0009] Preferably, the dynamically adjusting the maintenance cycle by using reinforcement learning based on the prediction of the remaining life of the equipment includes: Obtain the average daily usage time and average daily usage intensity of the embedded computer used by the user; Simulate the daily operation of high-precision digital models in the virtual world and accelerate the flow of time in virtual time; If the data simulation drive cannot use the software normally, it is determined that the data simulation drive is damaged; Obtain the operating time of the high-precision digital model from operation to damage, and record it as the remaining life; Obtain daily user usage, update the average daily usage time and usage intensity of the embedded computer, and obtain the remaining update life of the embedded computer; Based on the remaining life of the embedded computer, the initial maintenance cycle of the embedded computer is calculated using a cycle calculation formula; During the initial maintenance cycle, determine whether the remaining life of the update is greater than the remaining life of the update on the previous day. If so, add one day to the initial maintenance cycle; if not, subtract one day from the initial maintenance cycle. The cycle calculation formula is: , Where T is the initial maintenance cycle, A is the remaining life of the embedded computer, and B is the average maintenance times of the embedded computer.

[0010] Preferably, the human-machine collaborative decision-making module deploys a lightweight CNN model on the device side to achieve local real-time analysis of the vibration spectrum, and the machine evaluation includes: Deploy the CNN model on embedded computing to obtain the vibration signal of the embedded computer; The CNN model is used to analyze the vibration signal of the embedded computer in real time and convert it into an image, which is recorded as a real-time signal; Using artificial intelligence to analyze and store vibration images of at least one embedded computer operating normally, complete machine learning, and obtain normal signals; Artificial intelligence analysis compares the real-time signal with the normal signal to obtain the similarity between the real-time signal and the normal signal, and the similarity is expressed as a percentage; The difference between 1 and the similarity is recorded as the machine evaluation; The output of the CNN model is a feature vector processed by convolutional layers, pooling layers and fully connected layers.

[0011] Preferably, the human evaluation based on the high-precision digital model comprises: Obtain chip information, hardware information, and software operation status of high-precision digital models; Invite at least one expert to analyze chip information, hardware information, and software operation status to obtain an evaluation of the embedded computer; Analyze the authority and academic achievements of at least one expert, and use the analytic hierarchy process to obtain the analysis weight of each expert; The sum of the products of at least one expert's evaluation and the analysis weight is calculated to obtain the human evaluation.

[0012] Preferably, the determination is made whether the machine evaluation is greater than a preset value, and if so, the human evaluation is calculated; The initial weight of the machine evaluation is equal to the initial weight of the human evaluation; The product of the machine evaluation plus 1 and the initial weight of the machine evaluation is obtained to obtain the weight of the machine evaluation; The difference between 1 and the weight evaluated by the machine is recorded as the weight evaluated by humans; The sum of the weight of the machine evaluation and the product of the machine evaluation and the weight of the human evaluation and the product of the human evaluation is the human-machine collaborative evaluation; Get the number and priority of components maintained by the embedded computer; Based on the number and priority of the maintenance components of the embedded computer, it is divided into several maintenance levels; The human-machine collaboration assessment is divided into several intervals, each of which corresponds to a maintenance level from small to large; Determine the interval to which the human-machine collaborative assessment belongs and perform the corresponding maintenance level; The steps for obtaining the preset value are as follows: Counting the number of times at least one machine-assessed embedded computer fails; The minimum machine evaluation with the number of failures greater than 1 is selected as the preset value.

[0013] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes an embedded computer predictive maintenance control system based on digital twins, which uses digital twins to build a high-precision digital model of the embedded computer and builds a one-to-one restored model in the virtual world, which helps to achieve cost-free simulation operation and improve the accuracy of prediction. The fault anomaly detection module can promptly detect faults of the embedded computer to avoid the situation where the embedded computer fails during maintenance and maintenance cannot be performed. The predictive equipment maintenance module uses a high-precision digital model to determine the predicted life of the embedded computer, thereby dynamically adjusting the maintenance cycle, solving the problems of increased failure probability of embedded computers with too short maintenance cycles and excessively high costs for too long maintenance cycles. The human-computer collaborative decision-making module integrates accurate computer predictions and human experience predictions, greatly improving the accuracy of fault probability predictions. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flow chart of the embedded computer predictive maintenance control system based on digital twin of the present invention; Figure 2 A schematic diagram of the process of setting sensors to monitor the temperature and vibration data of an embedded computer in real time, uploading the data to the cloud, and using digital twins to build a high-precision digital model; Figure 3 A schematic diagram of the process of evaluating the junction temperature change of a chip and setting a junction temperature warning critical point according to the present invention; Figure 4 A schematic diagram of a process of extracting the energy entropy value of a vibration signal by using wavelet packet decomposition, determining whether the energy entropy value exceeds a preset value, and if so, determining whether an abnormality occurs in an embedded computer; Figure 5 This is a schematic diagram of a process of dynamically adjusting the maintenance cycle using reinforcement learning based on equipment remaining life prediction of the present invention; Figure 6 A lightweight CNN model is deployed on the device side for the human-machine collaborative decision-making module of the present invention to realize local real-time analysis of the vibration spectrum and obtain a flow chart of machine evaluation; Figure 7 A schematic diagram of a process of obtaining a human evaluation based on a high-precision digital model of the present invention; Figure 8 The present invention is a flowchart for determining whether the machine evaluation is greater than a preset value, and if so, calculating the human evaluation. DETAILED DESCRIPTION

[0015] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0016] Reference Figure 1 As shown, the embedded computer predictive maintenance control system based on digital twins includes: A data fusion modeling module, wherein the data fusion modeling module sets sensors to monitor the temperature and vibration data of the embedded computer in real time, uploads the data to the cloud, and uses the digital twin to build a high-precision digital model; A fault anomaly detection module, which evaluates the junction temperature change of the chip based on a high-precision digital model, sets a junction temperature warning critical point, and issues a warning if the junction temperature of the chip exceeds the warning critical point. Based on the high-precision digital model, the module uses wavelet packet decomposition to extract the energy entropy value of the vibration signal to determine whether the energy entropy value exceeds a preset value. If so, it is determined that an abnormality has occurred in the embedded computer; A predictive equipment maintenance module, which uses reinforcement learning to dynamically adjust the maintenance cycle based on the prediction of the remaining life of the equipment; A human-machine collaborative decision-making module deploys a lightweight CNN model on the device side to achieve local real-time analysis of the vibration spectrum, obtain a machine evaluation, obtain a human evaluation based on a high-precision digital model, merge the machine evaluation and the human evaluation to obtain a human-machine collaborative evaluation, and perform maintenance control.

[0017] It can be explained that using digital twins to build a high-precision digital model of the embedded computer and building a one-to-one restored model in the virtual world can help achieve cost-free simulation operation and improve the accuracy of prediction. The fault anomaly detection module can detect the fault of the embedded computer in time to avoid the situation where the embedded computer fails during maintenance and maintenance cannot be carried out. The predictive equipment maintenance module uses a high-precision digital model to determine the predicted life of the embedded computer, so as to dynamically adjust the maintenance cycle, solving the problems of increased failure probability of embedded computers with too short maintenance cycles and excessively high costs for too long maintenance cycles. The human-computer collaborative decision-making module integrates accurate computer predictions and human experience predictions, greatly improving the accuracy of fault probability predictions.

[0018] Reference Figure 2 As shown, sensors are set to monitor the temperature and vibration data of the embedded computer in real time, and the data is uploaded to the cloud. The digital twin is used to build a high-precision digital model, including: Obtain the external dimensions of the embedded computer and construct a three-dimensional model of the embedded computer; Obtain the hardware component model, usage time and software composition of the embedded computer, and build a preliminary high-precision digital model of the embedded computer based on the three-dimensional model of the embedded computer; Obtain the temperature, vibration, and electrical parameter data of the embedded computer when it is running, and record them as real parameters; Use the preliminary high-precision digital model to simulate the normal operation of the embedded computer, and record the temperature, vibration, and electrical parameter data of the high-precision digital model as simulation parameters; Obtaining the normal fluctuation range of each parameter when the embedded computer is operating normally, calculating the length of the normal fluctuation range of each parameter, and recording half of the length of the normal fluctuation range of each parameter as the fluctuation preset value of each parameter; Compare the real parameters with the simulated parameters, calculate the difference between the real parameters and the simulated parameters, list each parameter as the target parameter one by one, and determine whether the difference of the target parameter is greater than the preset value of the fluctuation of the target parameter. If so, record the target parameter as the influencing parameter; Adjust the hardware model and usage time of the high-precision digital model, as well as the software composition, and record the changes in various parameters of the high-precision digital model; The hardware model and usage time of the high-precision digital model are adjusted into numbers according to the release year of the model and the length of usage time, which are recorded as independent variable parameters, and the degree of change of each parameter is quantified into numbers, which are recorded as dependent variable parameters; Based on the changes in various parameters of the high-precision digital model, the Pearson correlation coefficient between the independent variable parameters and the dependent variable parameters is calculated to obtain the correlation matrix between the independent variable parameters and the dependent variable parameters; Compare the difference between the influencing parameter and the true parameter, and determine the adjustment degree of the independent variable parameter corresponding to the influencing parameter based on the product of the Pearson coefficient of the correlation matrix of the influencing parameter and the difference; Summarize the adjustment degree of the independent variable parameter corresponding to at least one influencing parameter to obtain adjustment details of hardware and software; Based on the required hardware and software adjustment details, the preliminary high-precision digital model is adjusted to obtain a high-precision digital model; The Pearson correlation coefficient is a statistic used to measure the degree of linear correlation between two continuous variables, and its value range is -1 to 1.

[0019] It can be understood that the high-precision digital model is constructed using temperature, vibration, and electrical parameter data. The temperature, vibration, and electrical parameter data can reflect the operation of the embedded computer and help to build a more accurate high-precision digital model. Based on the comparison of the high-precision digital model and the temperature, vibration, and electrical parameter data of the normal operation of the embedded computer, the hardware and software configuration of the high-precision digital model is adjusted to simulate the operation of the embedded computer with a more accurate high-precision digital model to improve the accuracy of predictive maintenance. The elements of each row in the correlation matrix are composed of the Pearson correlation coefficient of the degree of change of each parameter with one variable in the hardware model and the usage time according to the year of release of the model and the length of usage time, reflecting the degree of linear correlation between the two, and providing a reference for adjusting the hardware model and the usage time according to the year of release of the model and the length of usage time to make the temperature, vibration, and electrical parameter data meet the standards.

[0020] Reference Figure 3 As shown in the figure, the junction temperature change of the chip is evaluated and the junction temperature warning critical point is set, including: Get the temperature and air flow of the environment where the embedded computer is located; The average temperature at the chip node in the high-precision digital model is obtained by using the data in the high-precision digital model, which is recorded as the junction temperature of the chip; Obtain the time and load of common software used by users on embedded computers; Get the junction temperature when the user uses all common software at the same time, and record it as the maximum junction temperature; Evaluate the chip aging rate and performance when the maximum junction temperature is exceeded, and plot the change in junction temperature with the chip aging rate and performance; Find the point with the smallest derivative on the image and record the corresponding junction temperature as the warning critical point.

[0021] It is understandable that the junction temperature obtained by the high-precision digital model built using digital twins is simpler and more accurate than other methods. Maintaining the chip junction temperature within the normal range is an important way to ensure the normal power operation of the equipment and increase the service life of the equipment. If the junction temperature warning critical point is too high, it will accelerate the aging of the equipment and affect the operation rate of the equipment. If the junction temperature warning critical point is too low, it will be not conducive to the normal operation of the equipment. Therefore, it is very important to set an appropriate chip junction temperature warning critical point.

[0022] Reference Figure 4 As shown, wavelet packet decomposition is used to extract the energy entropy value of the vibration signal, and it is determined whether the energy entropy value exceeds a preset value. If so, it is determined that an abnormality occurs in the embedded computer, including: Obtain vibration signals of high-precision digital models; Decomposing the vibration signal by using wavelet packets to obtain at least one divided frequency segment; Determine whether the decomposition layer number of the vibration signal decomposed by the wavelet packet belongs to the standard decomposition layer number interval. If yes, calculate the energy entropy value of the vibration signal using the entropy value calculation formula. If no, obtain the vibration signal of the high-precision digital model again and decompose the vibration signal using the wavelet packet. Determine whether the energy entropy value exceeds a preset value, and if so, determine that an abnormality occurs in the embedded computer; The entropy value calculation formula is: , In the formula, H is the entropy energy, is the ratio of the number of frequency bands with the frequency of the ith frequency band to the number of all frequency bands, and n is the number of frequency bands after the vibration signal wavelet packet decomposition; The wavelet packet decomposition is to pass the signal through a series of filters with different center frequencies but the same bandwidth to obtain a signal of at least one frequency band; The steps for obtaining the preset value and the standard decomposition layer number interval are as follows: Obtain at least one vibration frequency during at least one time period when the embedded computer is operating normally; Decomposing at least one vibration frequency by using wavelet packets to obtain the decomposition layer number and energy entropy value of at least one vibration frequency; Draw a bar chart of the number of decomposition layers of vibration frequency and energy entropy value; Based on the bar chart, find out the decomposition layer interval corresponding to the concentrated distribution of energy entropy values, record it as the standard decomposition layer interval, and record the average entropy value corresponding to the concentrated distribution of energy entropy values ​​as the preset value.

[0023] It can be understood that the vibration signal of the embedded computer is decomposed using wavelet packets to obtain vibration frequencies of different frequency bands, and then the entropy calculation formula is used to calculate the energy entropy values ​​of different vibration signals after decomposition. Since the vibration signal of the embedded computer is regular during normal operation, if the energy entropy increase of the vibration signal is too large, it means that the embedded computer may have a fault. If the decomposition layer number of the vibration signal of the wavelet packet is too large or too small, it will affect the accuracy of the energy entropy value, so it is necessary to determine the range of the decomposition layer number.

[0024] Reference Figure 5 As shown in the figure, based on the prediction of the remaining life of the equipment, the maintenance cycle is dynamically adjusted using reinforcement learning, including: Obtain the average daily usage time and average daily usage intensity of the embedded computer used by the user; Simulate the daily operation of high-precision digital models in the virtual world and accelerate the flow of time in virtual time; If the data simulation drive cannot use the software normally, it is determined that the data simulation drive is damaged; Obtain the operating time of the high-precision digital model from operation to damage, and record it as the remaining life; Obtain daily user usage, update the average daily usage time and usage intensity of the embedded computer, and obtain the remaining update life of the embedded computer; Based on the remaining life of the embedded computer, the initial maintenance cycle of the embedded computer is calculated using a cycle calculation formula; During the initial maintenance cycle, determine whether the remaining life of the update is greater than the remaining life of the update on the previous day. If so, add one day to the initial maintenance cycle; if not, subtract one day from the initial maintenance cycle. The cycle calculation formula is: , Where T is the initial maintenance cycle, A is the remaining life of the embedded computer, and B is the average maintenance times of the embedded computer.

[0025] It can be explained that obtaining the remaining life of the embedded computer based on a high-precision digital model and adjusting the maintenance cycle in real time based on the remaining life of the embedded computer solves the problems of increased possibility of failure of embedded computers with too short maintenance cycles and too high costs due to too long maintenance cycles. Incorporating the average maintenance times of embedded computers into the calculation of the maintenance cycle of embedded computers can reduce the inconvenience caused by too large or too small maintenance times with the universality of statistical data.

[0026] Reference Figure 6 As shown in the figure, the human-machine collaborative decision-making module deploys a lightweight CNN model on the device side to achieve local real-time analysis of the vibration spectrum, and the machine evaluation includes: Deploy the CNN model on embedded computing to obtain the vibration signal of the embedded computer; The CNN model is used to analyze the vibration signal of the embedded computer in real time and convert it into an image, which is recorded as a real-time signal; Using artificial intelligence to analyze and store vibration images of at least one embedded computer operating normally, complete machine learning, and obtain normal signals; Artificial intelligence analysis compares the real-time signal with the normal signal to obtain the similarity between the real-time signal and the normal signal, and the similarity is expressed as a percentage; The difference between 1 and the similarity is recorded as the machine evaluation; The output of the CNN model is a feature vector processed by convolutional layers, pooling layers and fully connected layers.

[0027] What can be explained is that the powerful computing power of artificial intelligence is used to compare the similarity between the real-time signal and the normal signal of the vibration image, which fully utilizes the advantages of artificial intelligence. A normally operating embedded computer has a certain vibration frequency. The similarity between the real-time signal and the normal signal is used to represent the machine evaluation, which can improve the accuracy of the machine evaluation.

[0028] Reference Figure 7 As shown, based on the high-precision digital model, the human assessment includes: Obtain chip information, hardware information, and software operation status of high-precision digital models; Invite at least one expert to analyze chip information, hardware information, and software operation status to obtain an evaluation of the embedded computer; Analyze the authority and academic achievements of at least one expert, and use the analytic hierarchy process to obtain the analysis weight of each expert; The sum of the products of at least one expert's evaluation and the analysis weight is calculated to obtain the human evaluation.

[0029] It can be explained that by aggregating the evaluations of multiple experts and assigning different weights to each expert based on the expert's authority and academic achievements, the situation of individual expert misjudgment is avoided and the accuracy of manual evaluation is greatly improved. The advantage of manual evaluation is that the possibility of fault is judged based on the expert's rich experience.

[0030] Reference Figure 8 As shown, it is determined whether the machine evaluation is greater than the preset value. If so, the human evaluation is calculated; The initial weight of the machine evaluation is equal to the initial weight of the human evaluation; The product of the machine evaluation plus 1 and the initial weight of the machine evaluation is obtained to obtain the weight of the machine evaluation; The difference between 1 and the weight evaluated by the machine is recorded as the weight evaluated by humans; The sum of the weight of the machine evaluation and the product of the machine evaluation and the weight of the human evaluation and the product of the human evaluation is the human-machine collaborative evaluation; Get the number and priority of components maintained by the embedded computer; Based on the number and priority of the maintenance components of the embedded computer, it is divided into several maintenance levels; The human-machine collaboration assessment is divided into several intervals, each of which corresponds to a maintenance level from small to large; Determine the interval to which the human-machine collaborative assessment belongs and perform the corresponding maintenance level; The steps for obtaining the preset value are as follows: Counting the number of times at least one machine-assessed embedded computer fails; The minimum machine evaluation with the number of failures greater than 1 is selected as the preset value.

[0031] It can be explained that human-machine collaboration combines the accurate numbers obtained by the powerful computing power of machine evaluation with the rich expert experience of human evaluation, greatly improving the accuracy of prediction of the possibility of failure of embedded mobile devices. It divides maintenance levels based on human-machine collaborative evaluation to avoid excessively high maintenance levels due to excessively low human-machine collaboration, which would cause waste of human resources.

[0032] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored, and when the computer-readable program is called, the above-mentioned embedded computer predictive maintenance control system based on digital twins is executed.

[0033] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0034] To sum up, the advantages of the present invention are: the present invention proposes an embedded computer predictive maintenance control system based on digital twins, uses digital twins to build a high-precision digital model of the embedded computer, and builds a one-to-one restored model in the virtual world, which is helpful to achieve cost-free simulation operation and improve the accuracy of prediction. The fault anomaly detection module can promptly detect the fault of the embedded computer to avoid the situation where the embedded computer fails during maintenance and maintenance cannot be carried out. The predictive equipment maintenance module uses a high-precision digital model to determine the predicted life of the embedded computer, thereby dynamically adjusting the maintenance cycle, solving the problems of increased failure probability of embedded computers with too short maintenance cycles and excessively high costs due to too long maintenance cycles. The human-computer collaborative decision-making module integrates accurate computer predictions and human experience predictions, greatly improving the accuracy of fault probability predictions.

[0035] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. An embedded computer predictive maintenance control system based on digital twins, characterized in that: include: A data fusion modeling module, wherein the data fusion modeling module sets sensors to monitor the temperature and vibration data of the embedded computer in real time, uploads the data to the cloud, and uses the digital twin to build a high-precision digital model; A fault anomaly detection module, which evaluates the junction temperature change of the chip based on a high-precision digital model, sets a junction temperature warning critical point, and issues a warning if the junction temperature of the chip exceeds the warning critical point. Based on the high-precision digital model, the module uses wavelet packet decomposition to extract the energy entropy value of the vibration signal to determine whether the energy entropy value exceeds a preset value. If so, it is determined that an abnormality has occurred in the embedded computer; A predictive equipment maintenance module, which uses reinforcement learning to dynamically adjust the maintenance cycle based on the prediction of the remaining life of the equipment; A human-machine collaborative decision-making module deploys a lightweight CNN model on the device side to achieve local real-time analysis of the vibration spectrum, obtain a machine evaluation, obtain a human evaluation based on a high-precision digital model, merge the machine evaluation and the human evaluation to obtain a human-machine collaborative evaluation, and perform maintenance control.

2. The embedded computer predictive maintenance control system based on digital twin according to claim 1 is characterized in that: The sensors are set to monitor the temperature and vibration data of the embedded computer in real time, upload the data to the cloud, and use the digital twin to build a high-precision digital model, including: Obtain the external dimensions of the embedded computer and construct a three-dimensional model of the embedded computer; Obtain the hardware component model, usage time and software composition of the embedded computer, and build a preliminary high-precision digital model of the embedded computer based on the three-dimensional model of the embedded computer; Obtain the temperature, vibration, and electrical parameter data of the embedded computer when it is running, and record them as real parameters; Use the preliminary high-precision digital model to simulate the normal operation of the embedded computer, and record the temperature, vibration, and electrical parameter data of the high-precision digital model as simulation parameters; Obtaining the normal fluctuation range of each parameter when the embedded computer is operating normally, calculating the length of the normal fluctuation range of each parameter, and recording half of the length of the normal fluctuation range of each parameter as the fluctuation preset value of each parameter; Compare the real parameters with the simulated parameters, calculate the difference between the real parameters and the simulated parameters, list each parameter as the target parameter one by one, and determine whether the difference of the target parameter is greater than the preset value of the fluctuation of the target parameter. If so, record the target parameter as the influencing parameter; Adjust the hardware model and usage time of the high-precision digital model, as well as the software composition, and record the changes in various parameters of the high-precision digital model; The hardware model and usage time of the high-precision digital model are adjusted into numbers according to the release year of the model and the length of usage time, which are recorded as independent variable parameters, and the degree of change of each parameter is quantified into numbers, which are recorded as dependent variable parameters; Based on the changes in various parameters of the high-precision digital model, the Pearson correlation coefficient between the independent variable parameters and the dependent variable parameters is calculated to obtain the correlation matrix between the independent variable parameters and the dependent variable parameters; Compare the difference between the influencing parameter and the true parameter, and determine the adjustment degree of the independent variable parameter corresponding to the influencing parameter based on the product of the Pearson coefficient of the correlation matrix of the influencing parameter and the difference; Summarize the adjustment degree of the independent variable parameter corresponding to at least one influencing parameter to obtain adjustment details of hardware and software; Based on the required hardware and software adjustment details, the preliminary high-precision digital model is adjusted to obtain a high-precision digital model; The Pearson correlation coefficient is a statistic used to measure the degree of linear correlation between two continuous variables, and its value range is -1 to 1.

3. The embedded computer predictive maintenance control system based on digital twin according to claim 2 is characterized in that: The step of evaluating the change in junction temperature of the chip and setting the junction temperature warning critical point includes: Get the temperature and air flow of the environment where the embedded computer is located; The average temperature at the chip node in the high-precision digital model is obtained by using the data in the high-precision digital model, which is recorded as the junction temperature of the chip; Obtain the time and load of common software used by users on embedded computers; Get the junction temperature when the user uses all common software at the same time, and record it as the maximum junction temperature; Evaluate the chip aging rate and performance when the maximum junction temperature is exceeded, and plot the change in junction temperature with the chip aging rate and performance; Find the point with the smallest derivative on the image and record the corresponding junction temperature as the warning critical point.

4. The embedded computer predictive maintenance control system based on digital twin according to claim 3 is characterized in that: The method of extracting the energy entropy value of the vibration signal by wavelet packet decomposition and determining whether the energy entropy value exceeds a preset value, and if so, determining that an abnormality occurs in the embedded computer includes: Obtain vibration signals of high-precision digital models; Decomposing the vibration signal by using wavelet packets to obtain at least one divided frequency segment; Determine whether the decomposition layer number of the vibration signal decomposed by the wavelet packet belongs to the standard decomposition layer number interval. If yes, calculate the energy entropy value of the vibration signal using the entropy value calculation formula. If no, obtain the vibration signal of the high-precision digital model again and decompose the vibration signal using the wavelet packet. Determine whether the energy entropy value exceeds a preset value, and if so, determine that an abnormality occurs in the embedded computer; The entropy value calculation formula is: , In the formula, is the entropy energy, is the ratio of the number of frequency bands in the ith frequency band to the number of all frequency bands, is the number of frequency bands after wavelet packet decomposition of vibration signal; The wavelet packet decomposition is to pass the signal through a series of filters with different center frequencies but the same bandwidth to obtain a signal of at least one frequency band; The steps for obtaining the preset value and the standard decomposition layer number interval are as follows: Obtain at least one vibration frequency during at least one time period when the embedded computer is operating normally; Decomposing at least one vibration frequency by using wavelet packets to obtain the decomposition layer number and energy entropy value of at least one vibration frequency; Draw a bar chart of the number of decomposition layers of vibration frequency and energy entropy value; Based on the bar chart, find out the decomposition layer interval corresponding to the concentrated distribution of energy entropy values, record it as the standard decomposition layer interval, and record the average entropy value corresponding to the concentrated distribution of energy entropy values ​​as the preset value.

5. The embedded computer predictive maintenance control system based on digital twin according to claim 4 is characterized in that: The method of dynamically adjusting the maintenance cycle based on the prediction of the remaining life of the equipment by using reinforcement learning includes: Obtain the average daily usage time and average daily usage intensity of the embedded computer used by the user; Simulate the daily operation of high-precision digital models in the virtual world and accelerate the flow of time in virtual time; If the data simulation drive cannot use the software normally, it is determined that the data simulation drive is damaged; Obtain the operating time of the high-precision digital model from operation to damage, and record it as the remaining life; Obtain daily user usage, update the average daily usage time and usage intensity of the embedded computer, and obtain the remaining update life of the embedded computer; Based on the remaining life of the embedded computer, the initial maintenance cycle of the embedded computer is calculated using a cycle calculation formula; During the initial maintenance cycle, determine whether the remaining life of the update is greater than the remaining life of the update on the previous day. If so, add one day to the initial maintenance cycle; if not, subtract one day from the initial maintenance cycle. The cycle calculation formula is: , In the formula, For the initial maintenance cycle, is the remaining life of the embedded computer, is the average maintenance times of embedded computers.

6. The embedded computer predictive maintenance control system based on digital twin according to claim 5 is characterized in that: The human-machine collaborative decision-making module deploys a lightweight CNN model on the device side to achieve local real-time analysis of the vibration spectrum, and obtains machine evaluation including: Deploy the CNN model on embedded computing to obtain the vibration signal of the embedded computer; The CNN model is used to analyze the vibration signal of the embedded computer in real time and convert it into an image, which is recorded as a real-time signal; Using artificial intelligence to analyze and store vibration images of at least one embedded computer operating normally, complete machine learning, and obtain normal signals; Artificial intelligence analysis compares the real-time signal with the normal signal to obtain the similarity between the real-time signal and the normal signal, and the similarity is expressed as a percentage; The difference between 1 and the similarity is recorded as the machine evaluation; The output of the CNN model is a feature vector processed by convolutional layers, pooling layers and fully connected layers.

7. The embedded computer predictive maintenance control system based on digital twin according to claim 6 is characterized in that: The human evaluation based on the high-precision digital model includes: Obtain chip information, hardware information, and software operation status of high-precision digital models; Invite at least one expert to analyze chip information, hardware information, and software operation status to obtain an evaluation of the embedded computer; Analyze the authority and academic achievements of at least one expert, and use the analytic hierarchy process to obtain the analysis weight of each expert; The sum of the products of at least one expert's evaluation and the analysis weight is calculated to obtain the human evaluation.

8. The embedded computer predictive maintenance control system based on digital twin according to claim 7 is characterized in that: The combining of mechanized assessment and human assessment to obtain human-machine collaborative assessment and perform maintenance control includes: Determine whether the machine evaluation is greater than the preset value, and if so, calculate the human evaluation; The initial weight of the machine evaluation is equal to the initial weight of the human evaluation; The product of the machine evaluation plus 1 and the initial weight of the machine evaluation is obtained to obtain the weight of the machine evaluation; The difference between 1 and the weight evaluated by the machine is recorded as the weight evaluated by humans; The sum of the weight of the machine evaluation and the product of the machine evaluation and the weight of the human evaluation and the product of the human evaluation is the human-machine collaborative evaluation; Get the number and priority of components maintained by the embedded computer; Based on the number and priority of the maintenance components of the embedded computer, it is divided into several maintenance levels; The human-machine collaboration assessment is divided into several intervals, each of which corresponds to a maintenance level from small to large; Determine the interval to which the human-machine collaborative assessment belongs and perform the corresponding maintenance level; The steps for obtaining the preset value are as follows: Counting the number of times at least one machine-assessed embedded computer fails; The minimum machine evaluation with the number of failures greater than 1 is selected as the preset value.

Images