Fault monitoring method and system for nonwoven machinery production line based on artificial intelligence

Abstract

The present invention relates to the field of vibration data processing technology, and in particular to a method and system for monitoring the fault of a nonwoven machinery production line based on artificial intelligence. The method comprises: constructing an extreme triangle of a maximum point in a waveform diagram of a vibration signal when the nonwoven machinery is working, and obtaining a characteristic area and characteristic parameters of the maximum point; dividing the maximum point according to the characteristic parameters to obtain a high-density period and a low-density period; calculating the period characteristic parameters of the high-density period and classifying the high-density period to obtain a resonant period and a non-resonant period; interpolating and smoothing the maximum points of the low-density period and the resonant period in the waveform diagram respectively to obtain an optimized maximum point in the waveform diagram, and similarly obtaining an optimized minimum point; constructing an upper envelope and a lower envelope according to the optimized maximum point and minimum point in the waveform diagram to perform EMD decomposition to detect the fault of the nonwoven machinery. The present invention can enhance the reliability of fault detection for nonwoven machinery.

Description

Technical Field

[0001] The present invention relates to the technical field of vibration data processing, and in particular to a method and system for monitoring a nonwoven machinery production line fault based on artificial intelligence. Background Art

[0002] Non-woven machinery production lines are mainly used to produce non-woven fabrics. Non-woven fabrics are widely used in the medical, household goods and other industries due to their lightness, breathability and high strength. During the operation of the non-woven machinery production line, there are many fault problems, including equipment failure, electrical failure, cooling system failure, etc. Among them, equipment failure is an extremely important issue. It is directly related to product quality and production efficiency, and requires great attention in fault monitoring. Monitoring the vibration signal of mechanical equipment to evaluate the operating status of the equipment is currently the most common means, because its vibration signal contains a lot of information about the operating status of the mechanical equipment. The monitoring system can effectively reveal equipment failure problems by monitoring and analyzing the changing characteristics of the vibration signal, and promptly discover and notify relevant staff to take corresponding measures.

[0003] In the field of artificial intelligence, vibration signals are generally decomposed into intrinsic mode functions (IMFs) of different frequency components using EMD decomposition. The fault problems of mechanical equipment are effectively identified through the frequency, energy and other characteristics of each IMF. The traditional method of using empirical mode decomposition (EMD) requires extracting extreme points in the original signal to construct the envelope.

[0004] However, due to equipment failure, resonance, noise interference and other reasons, the distribution and quantity characteristics of the extreme points in the vibration signal are affected, which makes it difficult to ensure the accuracy of envelope construction, resulting in inaccurate intrinsic mode functions (IMFs) obtained by final decomposition. IMFs cannot correctly reflect the current equipment status, and the monitoring efficiency and accuracy are greatly reduced. Summary of the invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a nonwoven machinery production line fault monitoring method and system based on artificial intelligence. The technical solutions adopted are as follows:

[0006] In a first aspect, an embodiment of the present invention provides a nonwoven machinery production line fault monitoring method based on artificial intelligence, the method comprising:

[0007] Obtain the maximum point of the waveform of the vibration signal of the nonwoven machine when it is working, and use the first zero point on the left and right sides of the maximum point to construct an extreme triangle; use the area of ​​the extreme triangle to obtain the characteristic area of ​​the maximum point;

[0008] Obtain characteristic parameters of the maximum point according to the characteristic area of ​​the maximum point; divide the maximum point according to the characteristic parameters to obtain high-density time periods and low-density time periods;

[0009] Calculate the period characteristic parameters of the high-density period according to the characteristic parameters of the maximum point in the high-density period; classify the high-density period according to the period characteristic parameters to obtain the resonant period and the non-resonant period;

[0010] The maximum value points in the low-density period in the waveform are interpolated and the maximum value points in the resonance period are smoothed to obtain the optimized maximum value points in the waveform, and similarly, the optimized minimum value points in the waveform are obtained;

[0011] According to the optimized maximum and minimum points in the waveform, the upper and lower envelopes are constructed to perform EMD decomposition to detect nonwoven machinery faults.

[0012] Preferably, using the area of ​​the extreme value triangle to obtain the characteristic area of ​​the maximum value point includes:

[0013] The distance between the horizontal coordinates of a maximum point and the previous maximum point of the maximum point is obtained, which is recorded as the first distance; the area of ​​the extreme triangle of the maximum point is calculated, and the area of ​​the extreme triangle is compared with the first distance and normalized to obtain the characteristic area of ​​the maximum point.

[0014] Preferably, obtaining characteristic parameters of the maximum point according to the characteristic area of ​​the maximum point includes:

[0015] The difference between the characteristic area of ​​a maximum point and the mean of the characteristic areas of all maximum points is obtained, and the difference is input into the hyperbolic tangent function to obtain the characteristic parameters of the maximum point.

[0016] Preferably, the maximum value points are divided according to the characteristic parameters to obtain high-density time periods and low-density time periods, including:

[0017] Starting from the first maximum point, if the characteristic parameter of the first maximum point is greater than the first threshold, then start traversing until the characteristic parameter of the maximum point is less than or equal to the second threshold, then stop traversing, and the time period from the first maximum point to the end of the traversal is the high-density time period; if the characteristic parameter of the first maximum point is less than the third threshold, then start traversing until the characteristic parameter of the maximum point is greater than or equal to the fourth threshold, then stop traversing, and the time period from the first maximum point to the end of the traversal is the low-density time period; and so on, the next judgment and traversal are performed from the next maximum point in the high-density time period or the low-density time period, until all the maximum points in the waveform diagram participate in the traversal, and the high-density time period and the low-density time period are obtained.

[0018] Preferably, the calculation formula of the period characteristic parameter of the high-density period is specifically: ,in, represents the time period characteristic parameter of the kth high-density time period; and are the first weight coefficient and the second weight coefficient respectively; tanh represents the hyperbolic tangent function; represents the characteristic parameter of the jth maximum point in the kth high-density period; N represents the number of maximum points in the kth high-density period; exp{} represents an exponential function with a natural constant as the base; and They respectively represent the end time and start time of the kth high-density period.

[0019] Preferably, the high-density time periods are classified according to the time period characteristic parameters to obtain resonant time periods and non-resonant time periods, including:

[0020] A first partition interval and a second partition interval are set. If the period characteristic parameter of the high-density period is within the first partition interval, the high-density period is a resonance period; if the period characteristic parameter of the high-density period is within the second partition interval, the high-density period is a non-resonance period.

[0021] Preferably, interpolating the maximum value points of the low-density period in the waveform diagram includes:

[0022] Calculate the first distance of the horizontal coordinates between every two adjacent maximum points in the low-density period; obtain the average of the first distance and the distance of the horizontal coordinates between all two adjacent maximum points in the waveform diagram, recorded as the mean distance; the ratio of the first distance between two adjacent maximum points in the low-density period to the mean distance is the distance ratio; integer the distance ratio, and the difference between the integer distance ratio and the preset value is the number of interpolations required between two adjacent maximum points in the low-density period; uniformly interpolate between two adjacent maximum points in the low-density period according to the number of interpolations required.

[0023] Preferably, the maximum value point in the resonance period is smoothed, including:

[0024] A moving average filter is used to smooth out the maximum points within the resonance period.

[0025] In the second aspect, the present invention also provides an artificial intelligence-based non-woven machinery production line fault monitoring system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of a non-woven machinery production line fault monitoring method based on artificial intelligence are implemented.

[0026] The embodiments of the present invention have at least the following beneficial effects: the present invention obtains a waveform diagram of a vibration signal of a nonwoven machine when it is working, and then obtains a maximum point in the waveform diagram, and then constructs an extreme triangle of each maximum point according to the characteristics of resonance or fault in the vibration signal, which is used to analyze the characteristics of each maximum point, obtain the characteristic area of ​​the maximum point, and then obtain the characteristic parameters of the maximum point; further, the maximum point is divided into a high-density period and a low-density period using the characteristic parameters, and the high-density period is further analyzed to distinguish whether it is a resonance period caused by resonance, so as to facilitate subsequent processing thereof and reduce its influence on the fault detection of the nonwoven machine; finally, the maximum point of the low-density period in the waveform diagram is interpolated and the maximum point in the resonance period is smoothed to obtain the optimized maximum point in the waveform diagram, and the optimized minimum point is obtained according to this method, so as to achieve the purpose of optimizing the extreme point in the waveform diagram, obtain the optimal extreme point to construct the upper and lower envelopes, make the intrinsic mode functions (IMFs) obtained by EMD decomposition more accurate, and enhance the reliability of the fault monitoring of the nonwoven machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 A method flow chart of a method for monitoring faults in a nonwoven machinery production line based on artificial intelligence provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a non-woven machinery production line fault monitoring method and system based on artificial intelligence proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments may be combined in any suitable form.

[0030] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0031] The specific scheme of the nonwoven machinery production line fault monitoring method based on artificial intelligence provided by the present invention is described in detail below with reference to the accompanying drawings.

[0032] Exemplary methods:

[0033] The main application scenarios of the present invention are:

[0034] The present invention constructs an extreme value triangle by analyzing the extreme points of the vibration signal, obtains the characteristic area of ​​the extreme value triangle and the characteristic parameters of the extreme points, distinguishes the density time periods according to the characteristic parameters, and finally obtains the optimal extreme point to construct the envelope, so that the intrinsic mode functions (IMFs) obtained by EMD decomposition are more accurate and the reliability of fault monitoring is enhanced.

[0035] See also Figure 1 , which shows a method flow chart of a method for monitoring a nonwoven machinery production line fault based on artificial intelligence provided by an embodiment of the present invention, the method comprising the following steps:

[0036] Step S1, obtaining the maximum point of the waveform of the vibration signal of the nonwoven machine when it is working, constructing an extreme triangle using the first zero point on both sides of the maximum point; and obtaining the characteristic area of ​​the maximum point using the area of ​​the extreme triangle.

[0037] Obtain the vibration signal data generated by the nonwoven machinery of the production line, and use Python, MATLAB, etc. to generate a waveform graph based on the vibration signal data, where the X-axis is time and the Y-axis is amplitude. The waveform graph can clearly show the time domain characteristics of the vibration signal, including amplitude, periodicity, trend, and instantaneous changes, laying the foundation for subsequent feature analysis.

[0038] When using EMD to decompose the vibration signal of mechanical equipment, it is first necessary to construct the envelope through extreme points. However, due to interference such as equipment failure, resonance, and noise, the amplitude of the extreme points of the vibration signal will increase or decrease significantly, and the extreme points will appear in relatively dense or sparse areas, which seriously affects the accuracy of the envelope construction, causing the decomposed IMFs to fail to accurately reflect the equipment status. Therefore, by constructing an extreme triangle and calculating its characteristic area to comprehensively reflect the amplitude and frequency of the extreme points, and then obtaining the characteristic parameters of the extreme points through the characteristic area, different density areas are distinguished according to the characteristic parameters; then, due to the different natures of resonance and equipment failure, the time period characteristic parameters of the high-density period are obtained to distinguish the vibration signal changes caused by resonance and equipment failure, and the optimal extreme points are selected to construct the envelope.

[0039] Due to non-woven machinery failures, resonance phenomena and noise interference, the vibration signal may show characteristics that are inconsistent with normal conditions, such as relatively dense or sparse extreme points and large fluctuations in amplitude. The above characteristics will affect the accuracy of envelope construction and cause inaccurate IMFs obtained by EMD decomposition. Therefore, it is necessary to first screen the vibration signal to find extreme points that meet the above changes (before EMD decomposition, the upper and lower envelopes need to be constructed respectively through maximum and minimum values. Because their construction principles and means are the same, the subsequent operations of this solution take the construction of the upper envelope using the maximum point as an example).

[0040] Furthermore, the maximum value point in the waveform diagram is obtained, wherein the determination of the maximum value point can be determined by a derivative function, or by using an existing algorithm.

[0041] The scenario problem described in the core of the solution causes the extreme point distribution characteristics to be abnormal. Since the anomaly is mainly manifested in the obvious changes in the amplitude and frequency (i.e., density) of the extreme points, the extreme point and the first (nearest neighbor) zero-value point on the left and right sides (the intersection of the waveform and the X-axis) are connected to construct an extreme triangle, and the characteristic area of ​​the extreme triangle is obtained by comprehensively comparing the X-axis distance difference with the previous maximum point.

[0042] Due to the scene problem, the extreme points in the vibration signal have the characteristics of high amplitude and high concentration or low amplitude and low sparseness. The initial area of ​​the maximum point is directly proportional to the amplitude characteristics of the current maximum point, that is, the larger the area, the higher the amplitude; the smaller the area, the lower the amplitude. At the same time, the X-axis difference between adjacent maximum points can effectively reflect the frequency characteristics (that is, the density characteristics, because the faster the frequency, the denser the extreme points). The smaller the difference, the higher the frequency and denser the current maximum point. Conversely, the characteristic area can comprehensively reflect the characteristics of the above two types of maximum points, which facilitates subsequent feature analysis.

[0043] Further, the distance between the horizontal coordinates of a maximum point and the previous maximum point of the maximum point is obtained, recorded as the first distance; the area of ​​the extreme triangle of the maximum point is calculated, and the area of ​​the extreme triangle is compared with the first distance and normalized to obtain the characteristic area of ​​the maximum point; the calculation formula of the characteristic area of ​​the maximum point is specifically: , ,in, Represents the characteristic area of ​​the i-th maximum point; Represents the difference between the abscissa of the ith maximum point and the previous maximum point; The area of ​​the extreme triangle representing the i-th maximum point; and Respectively represent the horizontal coordinate values ​​of the first zero point on the left and right sides of the i-th maximum point; represents the ordinate of the ith maximum point; tanh represents the hyperbolic tangent function, and the output result is normalized and limited to (0,1). In this way, the characteristic area of ​​each maximum point can be obtained.

[0044] Step S2, obtaining characteristic parameters of the maximum point according to the characteristic area of ​​the maximum point; dividing the maximum point according to the characteristic parameters to obtain a high-density period and a low-density period.

[0045] Abnormal vibration signals caused by mechanical equipment failure and resonance phenomena will significantly increase the amplitude of extreme points in the vibration signal, and obvious extreme point concentration will appear; while abnormal vibration signals caused by noise interference, data acquisition errors, etc. will show low-amplitude vibration and relatively sparse distribution of extreme points. This feature will cause the characteristic area of ​​the extreme triangle to increase or decrease significantly, so the maximum points are screened through this change feature.

[0046] Furthermore, characteristic parameters of the maximum point are obtained according to the characteristic area of ​​the maximum point, the difference between the characteristic area of ​​a maximum point and the mean of the characteristic areas of all the maximum points is obtained, and the difference is input into the hyperbolic tangent function to obtain the characteristic parameters of the maximum point; the specific calculation formula of the characteristic parameters is: ,in, Represents the characteristic parameter of the i-th maximum point; Represents the characteristic area of ​​the i-th maximum point; represents the characteristic area of ​​the uth maximum point; n represents the number of maximum points in the waveform; tanh represents the hyperbolic tangent function, and the output result is limited to (-1,1); Represents the mean of the feature area.

[0047] Mechanical equipment failure and resonance phenomena will cause obvious extreme point concentration in the vibration signal, and the amplitude of the extreme point will increase significantly. Then the characteristic area of ​​the extreme triangle formed by the extreme points with such characteristics will be significantly higher than the average level. The closer it is to 1, the larger the amplitude of the i-th extreme point is and the higher the frequency is. You can set it according to different scenarios. This solution is for reference: When , the characteristic parameter of the maximum point is higher;

[0048] When there are background noise interference, data collection errors and other phenomena, the distribution of extreme points of the vibration signal will be relatively sparse, and the amplitude will be reduced. Then, when the extreme points with such characteristics appear, the characteristic area of ​​the extreme triangle they form will be significantly lower than the average level. The closer it is to -1, the smaller the amplitude of the i-th extreme point is, and the lower the frequency is. You can set it according to different scenarios. This solution is for reference: , the characteristic parameter of this maximum point is low.

[0049] Because the vibration signal of the equipment is an indicator of the vibration condition of the equipment, and the vibration of the machine is continuous, this means that when there is equipment failure, resonance, noise interference, data acquisition error and other phenomena, the extreme point change characteristics of the vibration signal will continue for a certain period of time. Based on this phenomenon, combined with the characteristic parameters of the maximum point, the maximum point is divided into high-density periods and low-density periods. The high-density period may be caused by resonance and failure, and the low-density period is caused by noise interference, data acquisition error and other phenomena.

[0050] Starting from the first maximum point, if the characteristic parameter of the first maximum point is greater than the first threshold, then start traversing until the characteristic parameter of the maximum point is less than or equal to the second threshold, then stop traversing, and the time period from the first maximum point to the end of the traversal is the high-density time period; if the characteristic parameter of the first maximum point is less than the third threshold, then start traversing until the characteristic parameter of the maximum point is greater than or equal to the fourth threshold, then stop traversing, and the time period from the first maximum point to the end of the traversal is the low-density time period; and so on, the next judgment and traversal are performed from the next maximum point in the high-density time period or the low-density time period, until all the maximum points in the waveform diagram participate in the traversal, and the high-density time period and the low-density time period are obtained.

[0051] It should be noted that the values ​​of the first threshold, the second threshold, the third threshold and the fourth threshold are 0.6, 0.2, -0.6 and -0.2 respectively, and the implementer can adjust them according to the actual situation. The characteristic parameters of the maximum value point in the high-density period are larger, and the characteristic parameters of the maximum value point in the low-density period are smaller.

[0052] Step S3, calculating the period characteristic parameters of the high-density period according to the characteristic parameters of the maximum point in the high-density period; and classifying the high-density period according to the period characteristic parameters to obtain the resonant period and the non-resonant period.

[0053] During high-density periods, the maximum points will be obviously concentrated due to equipment failure or resonance, accompanied by higher amplitudes. However, resonance itself is a physical phenomenon and does not belong to mechanical failure. The two need to be distinguished in fault monitoring.

[0054] The signal change caused by resonance lasts longer, and the amplitude of the maximum point is relatively stable; while the signal change caused by equipment failure lasts shorter, and the amplitude of the maximum point varies greatly. Therefore, the time period characteristic parameters of the high-density period are obtained through the above characteristics to distinguish the signal change caused by resonance from the equipment failure.

[0055] The period characteristic parameters of the high-density period are calculated according to the characteristic parameters of the maximum point in the high-density period. The calculation formula of the period characteristic parameters is as follows: ,in, represents the time period characteristic parameter of the kth high-density time period; and are the first weight coefficient and the second weight coefficient respectively; tanh represents the hyperbolic tangent function; represents the characteristic parameter of the jth maximum point in the kth high-density period; N represents the number of maximum points in the kth high-density period; exp{} represents an exponential function with a natural constant as the base; and They respectively represent the end time and start time of the kth high-density period.

[0056] represents the variance of the characteristic parameter of the maximum point in the kth high-density period, where and The value in the present invention is 0.5, and can be set according to different scenarios and signals; The closer it is to 0, the more stable the characteristic parameter change of the maximum point in the high-density period is. At the same time, the longer the high-density period lasts, the more it conforms to the signal change characteristics caused by the resonance phenomenon.

[0057] Furthermore, the distinction intervals are set, namely the first distinction interval and the second distinction interval, wherein the first distinction interval is (0, 0.4], and the second distinction interval is (0.4, 1). The implementer can determine the range of the first distinction interval and the second distinction interval according to the actual situation; if the time period characteristic parameter of the high-density time period is within the first distinction interval, the high-density time period is a resonance time period; if the time period characteristic parameter of the high-density time period is within the second distinction interval, the high-density time period is a non-resonance time period. The resonance time period is caused by the resonance phenomenon, and the non-resonance time period is caused by a mechanical failure.

[0058] Step S4, interpolating the maximum points in the low-density period in the waveform and smoothing the maximum points in the resonance period to obtain the optimized maximum points in the waveform, and similarly obtaining the optimized minimum points in the waveform.

[0059] Step S3 distinguishes different density time periods of extreme points. For low-density time periods, the maximum points of this part may be sparsely distributed due to noise interference, measurement errors, damping effects, etc., which leads to insufficient extreme points to construct the envelope. The IMFs decomposed by EMD cannot accurately reflect the essential characteristics of the signal, affecting the validity of subsequent analysis results. Therefore, for the maximum points of low-density time periods, the maximum points of this part of the time period are interpolated to supplement the missing extreme points; for high-density time periods, because these time periods are mainly caused by equipment failure and resonance, and resonance itself is a physical phenomenon, which is not equivalent to mechanical failure, a filter is used to smooth out the abnormal extreme points caused by the resonance phenomenon, so that IMFs can more focus on reflecting equipment failure problems.

[0060] Specifically, the maximum points of the low-density period in the waveform are interpolated. Specifically, the distance between the horizontal coordinates of every two adjacent maximum points in the low-density period is calculated, which is the first distance; the average of the first distance and the distance between the horizontal coordinates of all two adjacent maximum points in the waveform is obtained, which is recorded as the mean distance; the ratio of the first distance between two adjacent maximum points in the low-density period to the mean distance is the distance ratio; the distance ratio is integerized, and the difference between the integer distance ratio and the preset value is the number of interpolations required between two adjacent maximum points in the low-density period; uniform interpolation is performed between two adjacent maximum points in the low-density period according to the number of interpolations required. The calculation formula of the distance ratio is as follows: , where Q represents the distance ratio between two adjacent maximum points in the low-density period; Indicates the distance between the horizontal coordinates of two adjacent maximum points in the low-density period, that is, the first distance between two adjacent maximum points in the low-density period; It represents the difference between the horizontal coordinates of the ith maximum point and the previous maximum point in the waveform, that is, the distance between the horizontal coordinates of the ith maximum point and the previous maximum point in the waveform; n represents the number of maximum points in the waveform. The preset value is 1, (Q-1) is the number of interpolations required between two adjacent maximum points in the low-density period, and the number of interpolations required between each two adjacent maximum points may be the same or different. The distance between two adjacent maximum points is , starting from the left maximum point, / (Q-1) distance marks are marked as missing points, that is, interpolation is performed uniformly between two adjacent maximum points. Further, known data points are input, and a cubic spline interpolation model is created using SciPy's interpolation function to generate interpolation points, and interpolation is performed between every two adjacent maximum points.

[0061] For the maximum points within the resonance period, a moving average filter is used to smooth out these maximum points, so that the IMFs obtained after EMD decomposition can more specifically reflect equipment failure problems.

[0062] Thus, the optimized maximum point in the waveform graph can be obtained. Similarly, the optimized minimum point in the waveform graph can be obtained by performing the same operation on the minimum point in the waveform graph.

[0063] Step S5, constructing an upper envelope and a lower envelope according to the optimized maximum and minimum points in the waveform diagram to perform EMD decomposition to detect the fault of the nonwoven machinery.

[0064] After obtaining the optimized maximum and minimum points in the waveform in the above steps, a more accurate envelope is constructed by obtaining the optimal extreme points, and the filtered maximum and minimum points are connected using cubic spline difference or other methods to construct the upper and lower envelopes, and obtain the average envelope. Subtract the average envelope from the original signal to obtain the first IMF component. Then use the residual signal as the new input signal, repeat the above steps of constructing the upper and lower envelopes, and obtaining the average envelope, until the residual signal meets the preset stop condition or the number of iterations reaches the set value.

[0065] Furthermore, based on the frequency and energy distribution of each IMF, characteristic parameters related to the fault, such as frequency offset and energy mutation, are extracted; based on historical data or experience, thresholds for the characteristic parameters of each IMF are set, and the vibration data of the mechanical equipment is monitored online. When the threshold is exceeded, it is considered that the mechanical equipment may have a fault; when a fault is detected, the system notifies the relevant operators and records the fault information.

[0066] Example systems:

[0067] This embodiment provides a nonwoven machinery production line fault monitoring system based on artificial intelligence, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of a nonwoven machinery production line fault monitoring method based on artificial intelligence are implemented. Since a nonwoven machinery production line fault monitoring method based on artificial intelligence has been described in detail in the exemplary method, it will not be described in detail here.

[0068] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0069] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0070] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring faults in a nonwoven machinery production line based on artificial intelligence, characterized in that: The method includes: Obtain the maximum point of the waveform of the vibration signal of the nonwoven machine when it is working, and use the first zero point on the left and right sides of the maximum point to construct an extreme triangle; use the area of ​​the extreme triangle to obtain the characteristic area of ​​the maximum point; Obtain characteristic parameters of the maximum point according to the characteristic area of ​​the maximum point; divide the maximum point according to the characteristic parameters to obtain high-density time periods and low-density time periods; Calculate the period characteristic parameters of the high-density period according to the characteristic parameters of the maximum point in the high-density period; classify the high-density period according to the period characteristic parameters to obtain the resonant period and the non-resonant period; The maximum value points in the low-density period in the waveform are interpolated and the maximum value points in the resonance period are smoothed to obtain the optimized maximum value points in the waveform, and similarly, the optimized minimum value points in the waveform are obtained; According to the optimized maximum and minimum points in the waveform, the upper and lower envelopes are constructed to perform EMD decomposition to detect nonwoven machinery faults.

2. According to the artificial intelligence-based nonwoven machinery production line fault monitoring method of claim 1, it is characterized in that: The method of obtaining the characteristic area of ​​the maximum point by using the area of ​​the extreme value triangle includes: The distance between the horizontal coordinates of a maximum point and the previous maximum point of the maximum point is obtained, which is recorded as the first distance; the area of ​​the extreme triangle of the maximum point is calculated, and the area of ​​the extreme triangle is compared with the first distance and normalized to obtain the characteristic area of ​​the maximum point.

3. The method for monitoring nonwoven machinery production line faults based on artificial intelligence according to claim 1, characterized in that: The step of obtaining the characteristic parameters of the maximum point according to the characteristic area of ​​the maximum point includes: The difference between the characteristic area of ​​a maximum point and the mean of the characteristic areas of all maximum points is obtained, and the difference is input into the hyperbolic tangent function to obtain the characteristic parameters of the maximum point.

4. The method for monitoring nonwoven machinery production line faults based on artificial intelligence according to claim 1, characterized in that: The step of dividing the maximum value points according to the characteristic parameters to obtain high-density time periods and low-density time periods includes: Starting from the first maximum point, if the characteristic parameter of the first maximum point is greater than the first threshold, then start traversing until the characteristic parameter of the maximum point is less than or equal to the second threshold, then stop traversing, and the time period from the first maximum point to the end of the traversal is the high-density time period; if the characteristic parameter of the first maximum point is less than the third threshold, then start traversing until the characteristic parameter of the maximum point is greater than or equal to the fourth threshold, then stop traversing, and the time period from the first maximum point to the end of the traversal is the low-density time period; and so on, the next judgment and traversal are performed from the next maximum point in the high-density time period or the low-density time period, until all the maximum points in the waveform diagram participate in the traversal, and the high-density time period and the low-density time period are obtained.

5. The method for monitoring nonwoven machinery production line faults based on artificial intelligence according to claim 1, characterized in that: The calculation formula of the period characteristic parameter of the high-density period is specifically: ,in, represents the time period characteristic parameter of the kth high-density time period; and are the first weight coefficient and the second weight coefficient respectively; tanh represents the hyperbolic tangent function; represents the characteristic parameter of the jth maximum point in the kth high-density period; N represents the number of maximum points in the kth high-density period; exp{} represents an exponential function with a natural constant as the base; and They respectively represent the end time and start time of the kth high-density period.

6. The method for monitoring nonwoven machinery production line faults based on artificial intelligence according to claim 1, characterized in that: The method of classifying the high-density time periods according to the time period characteristic parameters to obtain the resonant time periods and the non-resonant time periods includes: A first partitioning interval and a second partitioning interval are set. If the period characteristic parameter of the high-density period is within the first partitioning interval, the high-density period is a resonance period; if the period characteristic parameter of the high-density period is within the second partitioning interval, the high-density period is a non-resonance period.

7. The method for monitoring nonwoven machinery production line faults based on artificial intelligence according to claim 1, characterized in that: The interpolation of the maximum value points of the low-density time period in the waveform diagram includes: Calculate the first distance of the horizontal coordinates between every two adjacent maximum points in the low-density period; obtain the average of the first distance and the distance of the horizontal coordinates between all two adjacent maximum points in the waveform diagram, recorded as the mean distance; the ratio of the first distance between two adjacent maximum points in the low-density period to the mean distance is the distance ratio; integer the distance ratio, and the difference between the integer distance ratio and the preset value is the number of interpolations required between two adjacent maximum points in the low-density period; uniformly interpolate between two adjacent maximum points in the low-density period according to the number of interpolations required.

8. The method for monitoring nonwoven machinery production line faults based on artificial intelligence according to claim 1, characterized in that: The smoothing of the maximum value point within the resonance period includes: A moving average filter is used to smooth out the maximum points within the resonance period.

9. A nonwoven machinery production line fault monitoring system based on artificial intelligence, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by the processor, the steps of the artificial intelligence-based nonwoven machinery production line fault monitoring method as described in any one of claims 1 to 8 are implemented.

Images