A FMS flexible intelligent production line safety early warning system based on data analysis

By setting up vibration sensors on the FMS flexible intelligent production line and using data analysis technology to screen and predict abnormal vibration sources, the problem of lack of equipment life prediction in the existing technology is solved, and early warning of abnormal vibration sources in the FMS flexible subsystem is achieved , improves production efficiency and equipment reliability.

CN118351668BActive Publication Date: 2024-11-08QINGDAO LEADING AUTOMATION EQUIP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410500421.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-24
Publication Date
2024-11-08
Estimated Expiration
2044-04-24

AI Technical Summary

Technical Problem

The prior art lacks methods for predicting the service life of components in FMS flexible subsystems, which makes it difficult to warning of equipment failures and affects production efficiency.

Method used

Design a FMS flexible intelligent production line safety warning system based on data analysis. By setting up a vibration sensor at each vibration source, collecting signals and using abnormal vibration source screening subsystem and life prediction subsystem, according to the discrete degree of vibration signal and distribution shape values ​​filter out abnormal vibration sources and alarm prompts when their lifespan is below the threshold.

Benefits of technology

Accurate screening and life prediction of abnormal vibration sources in the FMS flexible subsystem are realized, early warning is made, equipment failure is avoided, and production efficiency and equipment reliability are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118351668B_ABST
    Figure CN118351668B_ABST
Patent Text Reader

Abstract

The present invention discloses an FMS flexible intelligent production line safety early warning system based on data analysis, which belongs to the technical field of mechanical vibration measurement. In the present invention, a vibration sensor is arranged at each vibration source to collect vibration signals. The vibration signals are collected by a signal collection subsystem and transmitted to an abnormal vibration source screening subsystem to screen out abnormal vibration sources and normal vibration sources. Since there are multiple repeated FMS flexible subsystems on the FMS flexible intelligent production line, the life of the abnormal vibration source at the same position can be predicted based on the normal vibration source at the same position. The early warning subsystem is used to issue an alarm when the life is lower than a threshold value, thereby giving an early warning to the equipment that is about to fail.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical vibration measurement, and in particular to an FMS flexible intelligent production line safety early warning system based on data analysis. Background Art

[0002] FMS (Flexible Manufacturing System) is a highly automated, efficient and flexible manufacturing system. It can flexibly adjust the production process according to changes in production tasks and environment to meet the production needs of multiple varieties and small batches. FMS adopts a modular design, which is convenient for expansion and transformation according to different production needs. There are multiple FMS flexible subsystems set up on the FMS flexible intelligent production line. By adding or removing the FMS flexible subsystems, the FMS flexible intelligent production line can be expanded or transformed. With the widespread application of FMS, equipment will fail during use. Equipment failure is prone to cause production accidents and affect production efficiency. There is currently a lack of a method to predict the service life of components in the FMS flexible subsystem, so as to give early warning of equipment that is about to fail. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention provides an FMS flexible intelligent production line safety early warning system based on data analysis, which solves the problem of the lack of a method for predicting the service life of components in the FMS flexible subsystem.

[0004] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a safety early warning system for FMS flexible intelligent production line based on data analysis, comprising: FMS flexible intelligent production line, multiple FMS flexible subsystems, vibration sensors, signal collection subsystems, abnormal vibration source screening subsystems, life prediction subsystems and early warning subsystems;

[0005] A vibration sensor is provided at each vibration source on each FMS flexible subsystem of the FMS flexible intelligent production line;

[0006] The signal collection subsystem is used to collect vibration signals from various vibration sensors;

[0007] The abnormal vibration source screening subsystem is used to screen out abnormal vibration sources and normal vibration sources according to the discrete degree value and distribution shape value of the vibration signal at each vibration source;

[0008] The life prediction subsystem is used to predict the life of abnormal vibration sources at the same position on each FMS flexible subsystem according to the normal vibration sources at the same position;

[0009] The early warning subsystem is used to issue an alarm when the life of the abnormal vibration source is lower than a threshold value.

[0010] The beneficial effects of the present invention are as follows: in the present invention, a vibration sensor is arranged at each vibration source to collect vibration signals. The vibration signals are collected by the signal collection subsystem and transmitted to the abnormal vibration source screening subsystem to screen out abnormal vibration sources and normal vibration sources. Since there are multiple repeated FMS flexible subsystems on the FMS flexible intelligent production line, the life of the abnormal vibration source at the same position can be predicted based on the normal vibration source at the same position. The early warning subsystem is used to issue an alarm when the life is lower than the threshold, thereby giving an early warning to the equipment that is about to fail.

[0011] Further, the abnormal vibration source screening subsystem includes: a discrete degree value calculation unit, a distribution shape value calculation unit, a first classification unit, a second classification unit and a screening unit;

[0012] The discrete degree value calculation unit is used to calculate the discrete degree value of the vibration signal at each vibration source according to the vibration signal at each vibration source;

[0013] The distribution shape value calculation unit is used to calculate the distribution shape value of the vibration signal at each vibration source according to the vibration signal at each vibration source;

[0014] The first classification unit is used to classify the vibration sources at the same position according to the discrete degree values ​​of the vibration signals at each vibration source at the same position to obtain a first abnormal class;

[0015] The second classification unit is used to classify the vibration sources at the same position according to the distribution shape values ​​of the vibration signals at each vibration source at the same position to obtain a second abnormal class;

[0016] The screening unit is used to take the intersection of the first abnormal category and the second abnormal category to obtain the abnormal vibration source, and the other vibration sources except the abnormal vibration source are normal vibration sources.

[0017] The beneficial effects of the above-mentioned further scheme are as follows: in the present invention, two characteristic values ​​of the vibration signal are extracted: a discrete degree value and a distribution shape value; according to the discrete degree values ​​at the same position on multiple FMS flexible subsystems, the vibration sources are classified to obtain a first abnormal class; according to the distribution shape values ​​at the same position on multiple FMS flexible subsystems, the vibration sources are classified to obtain a second abnormal class; the intersection of the first abnormal class and the second abnormal class is taken to improve the accuracy of extracting abnormal vibration sources.

[0018] Furthermore, the expression of the discrete degree value calculation unit is:

[0019] ,

[0020] Among them, S is the discrete degree value, xi is the i-th vibration data in the vibration signal at the vibration source, N is the number of vibration data in the vibration signal, i is a positive integer, max{x i} is the maximum value of the vibration signal, min{x i} is to take the minimum value of the vibration signal.

[0021] The beneficial effect of the above further scheme is that the present invention reflects the fluctuation size of the vibration signal and the degree of dynamic change of the signal by dividing the square of the difference between each vibration data and the mean by the square of the difference between the maximum value and the minimum value in the signal.

[0022] Furthermore, the expression of the distribution shape value calculation unit is:

[0023] ,

[0024] Among them, R is the distribution shape value, x i is the i-th vibration data in the vibration signal at the vibration source, N is the number of vibration data in the vibration signal, and i is a positive integer.

[0025] The beneficial effect of the above further solution is that the distribution shape value is used to detect outliers or transient features in the signal. Since outliers usually cause the tail of the distribution to become heavier, a high distribution shape value can indicate the presence of outliers in the signal.

[0026] Furthermore, the first classification unit is specifically configured to: perform clustering processing on the discrete degree values ​​of the vibration signals at the vibration sources at the same position, and classify the corresponding vibration sources in the classes other than the class containing the most discrete degree values ​​as the first abnormal class;

[0027] The second classification unit is specifically configured to perform clustering processing on the distribution shape values ​​of vibration signals at each vibration source at the same position, and classify the corresponding vibration sources in the classes other than the class containing the most distribution shape values ​​into the second abnormal class.

[0028] Further, the life prediction subsystem includes: a mean sequence extraction unit, a gap sequence construction unit, a life prediction unit, a correction coefficient calculation unit and a life correction unit;

[0029] The mean sequence extraction unit is used to segment the vibration signals at the normal vibration source and the abnormal vibration source at the same position on each FMS flexible subsystem on the FMS flexible intelligent production line, and calculate the mean on each segment to construct the mean sequence of the normal vibration source and the mean sequence of the abnormal vibration source;

[0030] The gap sequence extraction unit is used to subtract the mean sequence of the abnormal vibration source from the mean sequence of each normal vibration source to obtain a plurality of gap sequences;

[0031] The life prediction unit is used to predict the life of the abnormal vibration source according to a plurality of gap sequences;

[0032] The correction coefficient calculation unit is used to calculate the correction coefficient according to the discrete degree value and distribution shape value of the abnormal vibration source and the discrete degree value and distribution shape value of the normal vibration source;

[0033] The life correction unit is used to correct the life of the abnormal vibration source by using a correction coefficient to obtain a corrected life of the abnormal vibration source.

[0034] The beneficial effects of the above further scheme are as follows: in the present invention, the vibration signal is segmented, and the mean is extracted from each segment, so as to obtain a mean sequence corresponding to the vibration signal, reducing the amount of data, and subtracting the mean sequence of the abnormal vibration source from the mean sequence of each normal vibration source to obtain a difference sequence, and using a life prediction unit to process multiple difference sequences to predict the life of the abnormal vibration source, and then using a correction coefficient to correct the life, so as to further improve the prediction accuracy of the life of the abnormal vibration source.

[0035] Further, the life prediction unit includes a plurality of fully connected layers and a life output layer;

[0036] The expression of each fully connected layer is:

[0037] ,

[0038] Among them, y n is the output of the nth fully connected layer, ω n,j is the jth weight of the nth fully connected layer, b n,j is the jth bias of the nth fully connected layer, x n,j is the jth element in the nth gap sequence, n and j are positive integers, and L is the number of elements in the gap sequence;

[0039] The expression of the lifetime output layer is:

[0040] , ,

[0041] Where, h is the life of the abnormal vibration source, is the output mean of multiple fully connected layers, ω is The weight of , e is a natural constant, and M is the number of fully connected layers.

[0042] The beneficial effect of the above further scheme is: in the present invention, a fully connected layer is used to predict the influence of a gap sequence on the life span, the mean is taken in the life span output layer, and the life span of the abnormal vibration source is predicted based on the mean.

[0043] Furthermore, the expression of the correction coefficient calculation unit is:

[0044] ,

[0045] Among them, γ is the correction coefficient, S anomaly is the discrete value of the abnormal vibration source, S m,Normal is the discrete value of the mth normal vibration source, R anomaly is the distribution shape value of the abnormal vibration source, R m,Normal is the distribution shape value of the mth normal vibration source, m is a positive integer, and K is the number of normal vibration sources.

[0046] Furthermore, the expression of the life correction unit is:

[0047] ,

[0048] Among them, h c Corrected life of abnormal vibration source.

[0049] The beneficial effect of the above further solution is: in the present invention, the distance between the correction coefficient and the discrete degree value is Distance from the distribution shape value about, in and The larger it is, the greater the difference between the characteristic value of the vibration signal of the abnormal vibration source and the characteristic value of the vibration signal of the normal vibration source, and the larger the correction coefficient, resulting in a shorter life after correction. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a system block diagram of an FMS flexible intelligent production line safety early warning system based on data analysis. DETAILED DESCRIPTION

[0051] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0052] like Figure 1 As shown, a safety early warning system for FMS flexible intelligent production line based on data analysis includes: FMS flexible intelligent production line, multiple FMS flexible subsystems, vibration sensors, signal collection subsystems, abnormal vibration source screening subsystems, life prediction subsystems and early warning subsystems;

[0053] A vibration sensor is provided at each vibration source on each FMS flexible subsystem of the FMS flexible intelligent production line;

[0054] The signal collection subsystem is used to collect vibration signals from various vibration sensors;

[0055] The abnormal vibration source screening subsystem is used to screen out abnormal vibration sources and normal vibration sources according to the discrete degree value and distribution shape value of the vibration signal at each vibration source;

[0056] The life prediction subsystem is used to predict the life of abnormal vibration sources at the same position on each FMS flexible subsystem according to the normal vibration sources at the same position;

[0057] The early warning subsystem is used to issue an alarm when the life of the abnormal vibration source is lower than a threshold value.

[0058] In the present invention, the vibration source is defined as: a component that generates vibration, and the vibration source can specifically be: various motor positions of the numerical control equipment on the FMS flexible subsystem, and the transmission equipment on the transportation system.

[0059] In the present invention, since there are multiple repeated FMS flexible subsystems on the FMS flexible intelligent production line, the structures of the multiple FMS flexible subsystems are the same. Therefore, the same position in the present invention refers to the same part of the multiple FMS flexible subsystems.

[0060] The abnormal vibration source screening subsystem comprises: a discrete degree value calculation unit, a distribution shape value calculation unit, a first classification unit, a second classification unit and a screening unit;

[0061] The discrete degree value calculation unit is used to calculate the discrete degree value of the vibration signal at each vibration source according to the vibration signal at each vibration source;

[0062] The distribution shape value calculation unit is used to calculate the distribution shape value of the vibration signal at each vibration source according to the vibration signal at each vibration source;

[0063] The first classification unit is used to classify the vibration sources at the same position according to the discrete degree values ​​of the vibration signals at each vibration source at the same position to obtain a first abnormal class;

[0064] The second classification unit is used to classify the vibration sources at the same position according to the distribution shape values ​​of the vibration signals at each vibration source at the same position to obtain a second abnormal class;

[0065] The screening unit is used to take the intersection of the first abnormal category and the second abnormal category to obtain the abnormal vibration source, and the other vibration sources except the abnormal vibration source are normal vibration sources.

[0066] The present invention extracts two characteristic values ​​of vibration signals: a discrete degree value and a distribution shape value. According to the discrete degree values ​​at the same position on multiple FMS flexible subsystems, the vibration sources are classified to obtain a first abnormal class. According to the distribution shape values ​​at the same position on multiple FMS flexible subsystems, the vibration sources are classified to obtain a second abnormal class. The intersection of the first abnormal class and the second abnormal class is taken to improve the accuracy of extracting abnormal vibration sources.

[0067] The expression of the discrete degree value calculation unit is:

[0068] ,

[0069] Among them, S is the discrete degree value, x i is the i-th vibration data in the vibration signal at the vibration source, N is the number of vibration data in the vibration signal, i is a positive integer, max{x i} is the maximum value of the vibration signal, min{x i} is to take the minimum value of the vibration signal.

[0070] The present invention reflects the fluctuation size of the vibration signal and the degree of dynamic change of the signal by dividing the square of the difference between each vibration data and the mean value by the square of the difference between the maximum value and the minimum value in the signal.

[0071] The expression of the distribution shape value calculation unit is:

[0072] ,

[0073] Among them, R is the distribution shape value, x i is the i-th vibration data in the vibration signal at the vibration source, N is the number of vibration data in the vibration signal, and i is a positive integer.

[0074] The distribution shape value is used to detect outliers or transient features in a signal. Since outliers usually cause the tail of a distribution to become heavier, a high distribution shape value can indicate the presence of outliers in a signal.

[0075] The first classification unit is specifically: clustering the discrete degree values ​​of the vibration signals at the vibration sources at the same position, and classifying the corresponding vibration sources in the classes other than the class with the most discrete degree values ​​as the first abnormal class;

[0076] The second classification unit is specifically configured to perform clustering processing on the distribution shape values ​​of vibration signals at each vibration source at the same position, and classify the corresponding vibration sources in the classes other than the class containing the most distribution shape values ​​into the second abnormal class.

[0077] In this embodiment, in the first classification unit, the specific process of clustering is: calculating the distance between each discrete degree value and other discrete degree values, classifying the discrete degree values ​​whose distance is lower than the distance threshold into one category, that is, classifying the discrete degree values ​​with similar values ​​into one category, and obtaining multiple categories.

[0078] In this embodiment, in the second classification unit, the specific process of clustering is: calculating the distance between each distribution shape value and other distribution shape values, and classifying the distribution shape values ​​whose distance is lower than the distance threshold into one category, that is, classifying similar distribution shape values ​​into one category, and obtaining multiple categories.

[0079] The life prediction subsystem includes: a mean sequence extraction unit, a gap sequence construction unit, a life prediction unit, a correction coefficient calculation unit and a life correction unit;

[0080] The mean sequence extraction unit is used to segment the vibration signals at the normal vibration source and the abnormal vibration source at the same position on each FMS flexible subsystem on the FMS flexible intelligent production line, and calculate the mean on each segment to construct the mean sequence of the normal vibration source and the mean sequence of the abnormal vibration source;

[0081] The gap sequence extraction unit is used to subtract the mean sequence of the abnormal vibration source from the mean sequence of each normal vibration source to obtain a plurality of gap sequences;

[0082] The life prediction unit is used to predict the life of the abnormal vibration source according to a plurality of gap sequences;

[0083] The correction coefficient calculation unit is used to calculate the correction coefficient according to the discrete degree value and distribution shape value of the abnormal vibration source and the discrete degree value and distribution shape value of the normal vibration source;

[0084] The life correction unit is used to correct the life of the abnormal vibration source by using a correction coefficient to obtain a corrected life of the abnormal vibration source.

[0085] In the present invention, the vibration signal is segmented, and the mean is extracted from each segment to obtain a mean sequence corresponding to the vibration signal, thereby reducing the amount of data. The mean sequence of the abnormal vibration source is subtracted from the mean sequence of each normal vibration source to obtain a difference sequence. A life prediction unit is used to process multiple difference sequences to predict the life of the abnormal vibration source, and then a correction coefficient is used to correct the life, thereby further improving the prediction accuracy of the life of the abnormal vibration source.

[0086] The life prediction unit includes a plurality of fully connected layers and a life output layer;

[0087] The expression of each fully connected layer is:

[0088] ,

[0089] Among them, y n is the output of the nth fully connected layer, ω n,j is the jth weight of the nth fully connected layer, b n,j is the jth bias of the nth fully connected layer, x n,j is the jth element in the nth gap sequence, n and j are positive integers, and L is the number of elements in the gap sequence;

[0090] The expression of the lifetime output layer is:

[0091] , ,

[0092] Where, h is the life of the abnormal vibration source, is the output mean of multiple fully connected layers, ω is The weight of , e is a natural constant, and M is the number of fully connected layers.

[0093] In the present invention, a fully connected layer is used to predict the influence of a gap sequence on the life span, the mean is taken in the life span output layer, and the life span of the abnormal vibration source is predicted based on the mean.

[0094] The expression of the correction coefficient calculation unit is:

[0095] ,

[0096] Among them, γ is the correction coefficient, S anomaly is the discrete value of the abnormal vibration source, S m,Normal is the discrete value of the mth normal vibration source, R anomaly is the distribution shape value of the abnormal vibration source, R m,Normal is the distribution shape value of the mth normal vibration source, m is a positive integer, and K is the number of normal vibration sources.

[0097] The expression of the life correction unit is:

[0098] ,

[0099] Among them, h c Corrected life of abnormal vibration source.

[0100] In the present invention, the distance between the correction coefficient and the discrete degree value Distance from the distribution shape value about, in and The larger it is, the greater the difference between the characteristic value of the vibration signal of the abnormal vibration source and the characteristic value of the vibration signal of the normal vibration source, and the larger the correction coefficient, resulting in a shorter life after correction.

[0101] In the present invention, a vibration sensor is arranged at each vibration source for collecting vibration signals. The vibration signals are collected by the signal collection subsystem and transmitted to the abnormal vibration source screening subsystem to screen out abnormal vibration sources and normal vibration sources. Since there are multiple repeated FMS flexible subsystems on the FMS flexible intelligent production line, the life of the abnormal vibration source at the same position can be predicted based on the normal vibration source at the same position. The early warning subsystem is used to issue an alarm when the life is lower than the threshold, thereby giving an early warning to the equipment that is about to fail.

[0102] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A FMS flexible intelligent production line safety early warning system based on data analysis, characterized in that: include: FMS flexible intelligent production line, multiple FMS flexible subsystems, vibration sensors, signal collection subsystem, abnormal vibration source screening subsystem, life prediction subsystem and early warning subsystem; A vibration sensor is provided at each vibration source on each FMS flexible subsystem of the FMS flexible intelligent production line; The signal collection subsystem is used to collect vibration signals from various vibration sensors; The abnormal vibration source screening subsystem is used to screen out abnormal vibration sources and normal vibration sources according to the discrete degree value and distribution shape value of the vibration signal at each vibration source; The life prediction subsystem is used to predict the life of abnormal vibration sources at the same position on each FMS flexible subsystem according to the normal vibration sources at the same position; The early warning subsystem is used to issue an alarm when the life of the abnormal vibration source is lower than a threshold value; The abnormal vibration source screening subsystem comprises: a discrete degree value calculation unit, a distribution shape value calculation unit, a first classification unit, a second classification unit and a screening unit; The discrete degree value calculation unit is used to calculate the discrete degree value of the vibration signal at each vibration source according to the vibration signal at each vibration source; The distribution shape value calculation unit is used to calculate the distribution shape value of the vibration signal at each vibration source according to the vibration signal at each vibration source; The first classification unit is used to classify the vibration sources at the same position according to the discrete degree values ​​of the vibration signals at each vibration source at the same position to obtain a first abnormal class; The second classification unit is used to classify the vibration sources at the same position according to the distribution shape values ​​of the vibration signals at each vibration source at the same position to obtain a second abnormal class; The screening unit is used to take the intersection of the first abnormal class and the second abnormal class to obtain the abnormal vibration source, and the other vibration sources except the abnormal vibration source are normal vibration sources; The expression of the discrete degree value calculation unit is: , Among them, S is the discrete degree value, x i is the i-th vibration data in the vibration signal at the vibration source, N is the number of vibration data in the vibration signal, i is a positive integer, max{x i } is the maximum value of the vibration signal, min{x i } is to take the minimum value of the vibration signal; The expression of the distribution shape value calculation unit is: , Among them, R is the distribution shape value, x i is the i-th vibration data in the vibration signal at the vibration source, N is the number of vibration data in the vibration signal, and i is a positive integer.

2. The FMS flexible intelligent production line safety early warning system based on data analysis according to claim 1 is characterized in that: The first classification unit is specifically: clustering the discrete degree values ​​of the vibration signals at the vibration sources at the same position, and classifying the corresponding vibration sources in the classes other than the class with the most discrete degree values ​​as the first abnormal class; The second classification unit is specifically configured to perform clustering processing on the distribution shape values ​​of vibration signals at each vibration source at the same position, and classify the corresponding vibration sources in the classes other than the class containing the most distribution shape values ​​into the second abnormal class.

3. The FMS flexible intelligent production line safety early warning system based on data analysis according to claim 1 is characterized in that: The life prediction subsystem includes: a mean sequence extraction unit, a gap sequence construction unit, a life prediction unit, a correction coefficient calculation unit and a life correction unit; The mean sequence extraction unit is used to segment the vibration signals at the normal vibration source and the abnormal vibration source at the same position on each FMS flexible subsystem on the FMS flexible intelligent production line, and calculate the mean on each segment to construct the mean sequence of the normal vibration source and the mean sequence of the abnormal vibration source; The gap sequence extraction unit is used to subtract the mean sequence of the abnormal vibration source from the mean sequence of each normal vibration source to obtain a plurality of gap sequences; The life prediction unit is used to predict the life of the abnormal vibration source according to a plurality of gap sequences; The correction coefficient calculation unit is used to calculate the correction coefficient according to the discrete degree value and distribution shape value of the abnormal vibration source and the discrete degree value and distribution shape value of the normal vibration source; The life correction unit is used to correct the life of the abnormal vibration source by using a correction coefficient to obtain a corrected life of the abnormal vibration source.

4. The FMS flexible intelligent production line safety early warning system based on data analysis according to claim 3 is characterized in that: The life prediction unit includes a plurality of fully connected layers and a life output layer; The expression of each fully connected layer is: , Among them, y n is the output of the nth fully connected layer, ω n,j is the jth weight of the nth fully connected layer, b n,j is the jth bias of the nth fully connected layer, x n,j is the jth element in the nth gap sequence, n and j are positive integers, and L is the number of elements in the gap sequence; The expression of the lifetime output layer is: , , Where, h is the life of the abnormal vibration source, is the output mean of multiple fully connected layers, ω is The weight of , e is a natural constant, and M is the number of fully connected layers.

5. The FMS flexible intelligent production line safety early warning system based on data analysis according to claim 4 is characterized in that: The expression of the correction coefficient calculation unit is: , Among them, γ is the correction coefficient, S anomaly is the discrete value of the abnormal vibration source, S m,Normal is the discrete value of the mth normal vibration source, R anomaly is the distribution shape value of the abnormal vibration source, R m,Normal is the distribution shape value of the mth normal vibration source, m is a positive integer, and K is the number of normal vibration sources.

6. The FMS flexible intelligent production line safety early warning system based on data analysis according to claim 5 is characterized in that: The expression of the life correction unit is: , Among them, h c Corrected life of abnormal vibration source.