Intelligent online control method and system for high-speed frame-type strander

Abstract

The invention relates to the technical field of general control regulation, in particular to an intelligent online control method and system for a high-speed frame-type strander, and the method comprises the steps: collecting a speed signal and a vibration signal, and decomposing a high-frequency sub-signal and a low-frequency sub-signal of the vibration signal; determining the trend consistency of the low-frequency sub-signal and the speed signal; performing time-frequency analysis on the high-frequency sub-signals, determining signal impact strength according to a peak value and kurtosis in a time domain, and calculating a variable coefficient of an amplitude value; in the frequency domain, the energy concentration degree is evaluated according to the entropy value of the energy, the variable coefficients of the frequency are calculated, the time-frequency consistency is determined according to the difference of the two variable coefficients, and the abnormal degree of the high-frequency sub-signals is quantified by combining the signal impact strength and the energy concentration degree; and determining a working condition abnormal index by integrating the trend consistency of the low-frequency sub-signal and the speed signal and the abnormal index of the high-frequency sub-signal, and controlling the high-speed frame-type strander. According to the method, the operation stability of the high-speed frame-type strander is improved.

Description

Technical Field

[0001] The present invention relates to the general field of control and regulation technology, and more particularly to an intelligent online control method and system for a high-speed frame stranding machine. Background Art

[0002] High-speed stranding machines are used in the production of high-end cables such as communications cables and high-voltage transmission lines. Their primary function is to precisely strand conductors through the synchronized rotation of multiple spools, improving production efficiency through high speeds. During high-speed operation, high-speed stranding machines require extremely high precision and stability to ensure uniform lay length and minimize wire breakage.

[0003] To achieve efficient and stable operation of high-speed frame stranding machines, modern intelligent control systems have been widely used. This type of control system integrates microprocessors, multi-source sensor networks and real-time control algorithms. By monitoring operating parameters in real time and performing analysis and processing, the high-speed frame stranding machine is precisely controlled based on the analysis and processing results to ensure stable operation of the equipment.

[0004] Currently, a Chinese patent document with publication number CN118151591B discloses a speed sensor-based intelligent control method for high-speed frame stranding machines. This technical solution installs a speed sensor on each spool to monitor the spool's rotational speed in real time and transmit the data to a central microprocessor. The microprocessor analyzes the received rotational speed data for anomalies, assessing the severity of the anomaly in real time and implementing control measures.

[0005] However, the aforementioned technical solution primarily relies on speed information for analysis, and this single-dimensional control strategy based on speed analysis has certain limitations. Specifically, due to the complex and variable nature of the spool's operation, speed information is low-frequency. This means that by the time a speed anomaly is detected, the fault has typically developed to a more serious state, often manifesting as large fluctuations or significant jumps. This makes it impossible to detect potential faults early and provide early warnings, which in turn affects the system's ability to respond to potential risks.

[0006] Therefore, there is an urgent need to develop a more accurate and sensitive intelligent online control method to identify potential faults and anomalies in advance, optimize the control strategy, and ensure the stability and efficiency of equipment during the production process. Summary of the Invention

[0007] In order to solve the problem that the existing intelligent control method of high-speed frame stranding machine based on speed sensor relies on low-frequency speed information, is difficult to detect early potential faults and cannot provide early warning, resulting in insufficient ability of the system to respond to potential risks, the present invention proposes an intelligent online control method and system for high-speed frame stranding machine.

[0008] In a first aspect, the present invention provides an intelligent online control method for a high-speed frame stranding machine, comprising: Collect the speed signal and vibration signal of the high-speed frame stranding machine during operation, and obtain the high-frequency and low-frequency sub-signals of the vibration signal through signal decomposition technology; For the low-frequency sub-signal, trend analysis is performed to determine the trend consistency between the low-frequency sub-signal and the velocity signal. For the high-frequency sub-signal, time-frequency analysis is performed, including: In the time domain, the signal impact strength of the high-frequency sub-signal is determined by the peak value and kurtosis of the high-frequency sub-signal, and a first coefficient of variation is calculated based on the amplitude of the high-frequency sub-signal at all sampling points; in the frequency domain, the energy concentration of the high-frequency sub-signal is evaluated based on the entropy value of the energy of the high-frequency sub-signal at all frequencies, and a second coefficient of variation is calculated based on all frequencies of the high-frequency sub-signal; and the time-frequency consistency of the high-frequency sub-signal is determined based on the difference between the first coefficient of variation and the second coefficient of variation; By fusing the signal's impact strength, energy concentration, and time-frequency consistency, the abnormality of the high-frequency sub-signal is determined; According to the trend consistency of the low-frequency sub-signal and the speed signal, as well as the abnormality degree of the high-frequency sub-signal, the abnormal working condition index of the high-speed frame stranding machine is comprehensively determined, and the high-speed frame stranding machine is controlled according to the abnormal working condition index.

[0009] This technical solution utilizes wavelet decomposition to decompose vibration signals into high-frequency and low-frequency sub-signals, enabling multi-source information analysis of the macroscopic motion state and microscopic characteristics of a high-speed stranding machine. Trend analysis is performed on the low-frequency sub-signals to determine their coordinated variation with the velocity signal. Since low-frequency vibration is primarily driven by the periodic motion of rotating components, the low-frequency sub-signals of the vibration signal are physically coupled to the velocity signal. This analysis allows for rapid quantification of the macroscopic stability of the machine and provides a preliminary assessment of any operating anomalies. For the high-frequency sub-signals, a joint time-frequency analysis is performed to analyze their detailed characteristics, comprehensively and accurately quantifying their degree of anomaly. Finally, an operating anomaly index is determined by combining the trend consistency between the low-frequency sub-signals and the velocity signal, as well as the degree of anomaly in the high-frequency sub-signals. This effectively integrates macroscopic stability and microscopic fault information, comprehensively assessing the operating condition of the high-speed stranding machine, and forming a complete logical chain from early fault identification to graded intelligent control. This enables timely detection of potential faults at an early stage, improving the operational stability of the high-speed stranding machine.

[0010] Preferably, the trend consistency of the low-frequency sub-signal and the speed signal is determined in the following manner: based on the amplitudes of the low-frequency sub-signal and the speed signal at all sampling points, the Pearson correlation coefficient of the low-frequency sub-signal and the speed signal is calculated, and the Pearson correlation coefficient is used as the global trend consistency; based on the difference between the amplitudes of the low-frequency sub-signal and the speed signal at each sampling point and the amplitudes at the previous sampling point, the direction of change of the amplitudes of the low-frequency sub-signal and the speed signal at each sampling point is determined; based on the direction of change of the amplitudes of the low-frequency sub-signal and the speed signal at the same sampling point, the local trend consistency of the amplitudes of the low-frequency sub-signal and the speed signal is determined; the global trend consistency and the local trend consistency indicators are fused to obtain the trend consistency of the low-frequency sub-signal and the speed signal.

[0011] This technical solution uses the Pearson correlation coefficient as the global trend consistency, which can reflect the linear correlation strength between the low-frequency sub-signals and the speed signal in the overall distribution, and embodies the macro-synergy; through the local trend consistency, it can accurately capture the synchronization of the low-frequency sub-signals and the speed signal in the micro-time series, make up for the defect of global analysis being insensitive to local fluctuations, and finally integrate global and local indicators, which not only retains the grasp of the overall operating trend, but also enhances the perception of local abnormal changes.

[0012] Preferably, the signal impact strength of the high-frequency sub-signal is determined based on the following method: A first deviation between the peak value of the high-frequency sub-signal and the reference peak value, and a second deviation between the kurtosis of the high-frequency sub-signal and the reference kurtosis are calculated, and a normalized value of the product of the first deviation and the second deviation is used as the signal impact strength of the high-frequency sub-signal.

[0013] This technical solution quantifies the impact intensity of high-frequency sub-signals by multiplying the peak deviation and the kurtosis deviation. Its design strictly conforms to the physical nature of mechanical failures: the first deviation directly reflects the magnitude of the signal's impact, and the second deviation reflects whether the signal's impact is dense. Since the impact caused by mechanical failures must be both high-intensity and high-density, the product of the two can specifically amplify this composite anomaly and more accurately distinguish between real fault impacts and false interference.

[0014] Preferably, the first coefficient of variation is the ratio of the standard deviation of the amplitude of the high-frequency sub-signal at all sampling points to the mean, and the second coefficient of variation is the ratio of the standard deviation of all frequencies of the high-frequency sub-signal to the mean.

[0015] Preferably, the time-frequency consistency of the high-frequency sub-signal is determined based on the following method: calculating the difference between the first coefficient of variation and the second coefficient of variation, and using the difference as a negative correlation indicator of the time-frequency consistency to quantify the time-frequency consistency.

[0016] Preferably, the energy concentration degree of the high-frequency sub-signal is determined based on the following method: normalizing the entropy value of the energy of the high-frequency sub-signal at all frequencies to obtain a normalized entropy value, and subtracting the normalized entropy value from 1 to obtain the value obtained as the energy concentration degree of the high-frequency sub-signal.

[0017] Preferably, the abnormality degree of the high-frequency sub-signal is determined based on the following method: the product of the signal impact intensity, energy concentration and time-frequency consistency is used as the abnormality degree of the high-frequency sub-signal.

[0018] This technical solution constructs a three-dimensional assessment system for the abnormality of high-frequency sub-signals. The signal impact intensity reflects the morphological characteristics of the fault impact, the energy concentration quantifies the significance of the fault characteristics, and the time-frequency consistency verifies the coordination of the fault characteristics in the time-frequency domain. It improves the sensitivity to early weak faults and makes the quantitative results of the abnormality degree more accurately reflect the actual status of the equipment.

[0019] Preferably, the abnormal operating condition index of the high-speed frame stranding machine is determined based on the following method: the trend consistency of the low-frequency sub-signal and the speed signal is used as the first parameter, the abnormal degree of the high-frequency sub-signal is used as the second parameter, the first parameter is converted into a negative correlation index of the abnormal operating condition index through a subtraction operation, and the product of the second parameter and the negative correlation index after performing a plus operation is used as the abnormal operating condition index.

[0020] This technical solution achieves the coordinated quantification of macro-stability and micro-abnormality through this negatively correlated product design, which not only avoids the misjudgment of early faults in a single dimension, but also can amplify potential early faults. By dynamically combining local abnormalities and overall stability through the product, the working condition assessment results are more accurate and reliable.

[0021] Preferably, the method for controlling the high-speed frame stranding machine is: setting a first-level threshold and a second-level threshold of the operating condition abnormality index; if the operating condition abnormality index is less than the first-level threshold, it is determined that the operating condition has not produced an early abnormality, and the high-speed frame stranding machine maintains the parameters corresponding to the current operating condition and continues to operate; if the operating condition abnormality index is greater than the first-level threshold and not greater than the second-level threshold, it is determined that the operating condition has produced a minor early fault, the parameters corresponding to the current operating condition are reduced according to a preset percentage and continue to operate, and the staff is prompted to continue to pay attention; if the operating condition abnormality index is greater than the second-level threshold, it is determined that the operating condition has produced an obvious early fault, the speed is immediately reduced to idle speed, and the staff is prompted to stop and check.

[0022] In a second aspect, the present invention also provides an intelligent online control system for a high-speed frame stranding machine, the intelligent online control system comprising a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement any step of the online control method.

[0023] The present invention has the following effects: This solution uses signal decomposition to separate vibration signals into high-frequency and low-frequency sub-signals, enabling multi-dimensional analysis of the equipment's macroscopic and microscopic characteristics. Trend consistency analysis of the low-frequency sub-signals quickly captures changes in the equipment's macroscopic stability and identifies any abnormalities. Joint time-frequency analysis of the high-frequency sub-signals accurately quantifies the signal's impact intensity, energy concentration, and time-frequency consistency, enabling multi-dimensional anomaly detection and improving early warning capabilities for faults and operational stability of high-speed frame stranding machines. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic flow chart of the method of the present invention; Figure 2 is a schematic flow chart of the method of step S3 of the present invention; Figure 3 It is a schematic flow chart of the method of step S4 of the present invention. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] The present invention provides an intelligent online control method for a high-speed frame stranding machine, such as Figure 1 As shown in , including: S1: Collect the speed signal and vibration signal of the high-speed frame stranding machine during operation.

[0027] High-precision vibration sensors and speed sensors are deployed on each spool of the high-speed frame stranding machine to collect the spool's vibration data and (rotation) speed data in real time at the same high frequency (5kHz). Since each moment is a sampling point, a vibration data and a rotation speed data are generated at each sampling point.

[0028] At any moment As the current moment, obtain the 5000 sampling points generated within 1 second (experience value) before the current moment, and use all the historical sampling points as the sampling points of the current moment. According to the sequence of vibration data and the sequence of speed data at all historical sampling points, the speed signal of the current moment is obtained. and vibration signals .

[0029] By precisely collecting the spool's speed and vibration signals, we construct a signal reflecting the equipment's real-time operating status. This process ensures the continuity and synchronization of the vibration and speed signals over time, while also balancing real-time performance with the integrity of historical features through a 1-second time window (5,000 sampling points), providing a reliable data foundation for subsequent analysis.

[0030] S2: Extract the high-frequency sub-signals and low-frequency sub-signals of the vibration signal through signal decomposition technology.

[0031] The speed signal reflects the core dynamic characteristics of the spool rotation in a high-speed frame stranding machine and is a key indicator of the equipment's macroscopic operational stability. It directly characterizes the spool's instantaneous speed, speed fluctuations, and long-term operating trends. The speed signal's fundamental frequency and harmonic components are closely related to the equipment's power output and the load status of the transmission system. Under normal operating conditions, the speed signal exhibits smooth, periodic fluctuations with small, regular amplitudes. However, when the spool mechanically jams, the drive motor output becomes unstable, or the transmission mechanism wears, the speed signal can exhibit abnormal fluctuations, such as sudden rises and falls and increased periodic deviations.

[0032] Vibration signals reflect various physical characteristics of the spool rotation in high-speed frame stranding machines. Low-frequency components primarily reflect macroscopic motion characteristics, such as periodic fluctuations in spool rotation and vibrations caused by overall imbalance. High-frequency components, on the other hand, focus on microscopic fault characteristics, such as impact between bearing rollers and rails and minute friction during gear meshing.

[0033] The low-frequency components of vibration signals are primarily influenced by macrostructural factors such as unbalanced forces in rotating parts and bearing concentricity deviations, and are closely correlated with the fundamental frequency of the velocity signal and its harmonic components. Under normal operating conditions, the low-frequency components of the vibration signal (low-frequency vibration) maintain stable energy transfer with the velocity signal through the mechanical structure, exhibiting a strong correlation. However, when equipment fails, such as due to bearing wear or loose couplings, changes in the stiffness and damping characteristics of the mechanical structure cause distortion in the energy transfer path, significantly reducing the correlation between the low-frequency components of the vibration signal and the velocity signal. Therefore, analyzing changes in the correlation between the low-frequency components of the vibration signal and the velocity signal provides critical indicative information for early fault identification and can effectively assist in determining the operating status of the equipment.

[0034] In signal decomposition technology, wavelet decomposition can effectively separate the different frequency components of the signal and can adapt to the multi-band characteristics of the vibration signal of the high-speed frame stranding machine.

[0035] Therefore, based on the above logical analysis of this step, the vibration signal Perform 5-layer (experience value) wavelet decomposition to obtain 6 wavelet coefficients, which are: 、 、 、 、 and , and each wavelet coefficient corresponds to a sub-signal, respectively 、 、 、 、 and .in, The corresponding sub-signal is the low-frequency sub-signal of the vibration signal, which covers the low-frequency band of the vibration signal. to The corresponding sub-signals are all high-frequency sub-signals of the vibration signal, covering the high-frequency band of the vibration signal.

[0036] S3: Determine the trend consistency of the low-frequency sub-signals of the velocity signal and the vibration signal.

[0037] As mentioned above, the velocity signal reflects the core dynamic characteristics of the spool rotation. Its fundamental frequency and harmonics are closely related to the load of the transmission system, while the low-frequency component of the vibration signal is not only affected by macroscopic structural factors, but is also closely related to the fundamental frequency and harmonics of the velocity signal. Under normal working conditions, the two maintain stable energy transfer through the mechanical structure and show a strong correlation.

[0038] Therefore, by judging the trend consistency of the low-frequency sub-signals of the speed signal and the vibration signal, the coordination of the equipment's macroscopic operating status can be verified from the signal feature level. If the trend consistency is stronger, it means that the dynamic characteristics of the spool rotation and the energy transfer path of the low-frequency vibration have not been significantly distorted, which indirectly confirms that the equipment's transmission system, mechanical structure and other macroscopic levels are operating normally; if the trend consistency is weak, it suggests that there may be early faults such as mechanical jamming, increased bearing concentricity deviation or loose coupling, resulting in obstructed or distorted energy transfer. Trend consistency provides a more direct and reliable judgment basis for equipment status assessment and fault warning, and is a further deepening of the correlation analysis between the two.

[0039] Therefore, based on the above logical analysis of this step, the trend consistency of the low-frequency sub-signals of the speed signal and the vibration signal is obtained as follows: Figure 2 As shown, including: S31: Use Pearson correlation coefficient to assess global trend consistency.

[0040] The Pearson correlation coefficient essentially measures the strength of the linear correlation between two variables, and its value range is [-1,1]. Here, the Pearson correlation analysis relies on the amplitude of the signal at all sampling points, which can objectively reflect the overall change trend correlation of the signal during the entire analysis period.

[0041] As previously mentioned, there is a stable energy transfer and characteristic correlation between the low-frequency sub-signal of the vibration signal and the velocity signal at the macro-operation level. This correlation exhibits significant trend consistency under normal operating conditions. Therefore, the Pearson correlation coefficient can be used to quantify this trend consistency, avoiding subjective judgment bias and making global correlation analysis more objective and operational.

[0042] Specifically, based on the amplitudes of the low-frequency sub-signal and the speed signal at all sampling points, the Pearson correlation coefficient of the low-frequency sub-signal and the speed signal is calculated as the global trend consistency of the low-frequency sub-signal and the speed signal.

[0043] Among them, the low-frequency sub-signal of the vibration signal and speed signal The Pearson correlation coefficient between them satisfies the following relationship:

[0044] In the formula, for and The Pearson correlation coefficient between and Global trend consistency. is the sequence number of the sampling point, is the total number of sampling points. The speed signal and vibration signal have the same time length and the same total number of sampling points. Speed ​​signal In the The amplitude at the sampling point, for The mean of the amplitudes at all sampling points, for In the The amplitude at the sampling point, for The mean of the amplitudes at all sampling points.

[0045] In the formula, The larger the value is, the stronger the correlation between the low-frequency sub-signals of the speed signal and the vibration signal is. , indicating that the low-frequency sub-signals of the speed signal and the vibration signal have a trend of increasing or decreasing at the same time. The more obvious the trend consistency is, the more likely the frame stranding machine is in normal steady-state operation. At this time, the low-frequency vibration is generated by the regular mechanical rotation caused by the rotation speed. The smaller the value, the weaker the correlation between the low-frequency sub-signals of the speed signal and the vibration signal, and the weaker the trend consistency. , indicating that there is no correlation at all and no trend consistency, it is more likely that there is a fault in the frame stranding machine, which generates additional movement that interferes with the normal energy transfer, making the low-frequency sub-signal of the vibration signal disordered, resulting in a significant decrease in its correlation with the speed signal, and the frame stranding machine is more likely to have an abnormality. , indicating that the low-frequency sub-signals of the speed signal and the vibration signal show a completely negative correlation and opposite changing trends. The more likely it is that there is a reverse force in the mechanical structure, suggesting that there is equipment failure.

[0046] To facilitate subsequent analysis, this step also maps the Pearson correlation coefficient to Specifically, based on the following conversion formula from Map to :

[0047] In this formula, for The converted value has a range of Within the interval, The larger it is, the closer it is to 1, indicating that the vibration signal is a low-frequency sub-signal. With speed signal The stronger the global trend consistency is, the closer it is to 0, indicating that the low-frequency sub-signal of the vibration signal With speed signal The weaker the global trend consistency.

[0048] Finally, As and Global trend consistency.

[0049] S32: Determine the local trend consistency based on the amplitude difference at adjacent sampling points.

[0050] First, from the perspective of equipment failure mechanism, the development of most mechanical abnormalities has the characteristic of gradually spreading from local abnormalities to global abnormalities. In the early stage of the failure, local abnormalities often only appear under specific working conditions (such as sudden speed changes and load fluctuations). As the failure intensifies, it will evolve into obvious global abnormalities.

[0051] The global trend consistency between the low-frequency sub-signal of the vibration signal and the velocity signal is calculated based on the overall characteristics of all sampling points. The result is essentially a macroscopic estimate of the signal correlation relationship. However, early equipment failures often manifest as localized, transient abnormal fluctuations (such as a momentary jam in the transmission mechanism or a short, irregular impact of a bearing roller). These anomalies typically occur at a few adjacent sampling points.

[0052] Therefore, by analyzing the amplitude differences at adjacent sampling points to measure local trend consistency, we can capture local dynamic distortions that are difficult to reflect through global analysis, and can also detect early abnormalities. Local trend consistency analysis is a necessary complement to global trend consistency analysis. By focusing on subtle correlations between signals at adjacent sampling points, it can improve the sensitivity of identifying early, localized faults. Furthermore, combined with global analysis, a combined macro- and micro-level trend consistency assessment system can be constructed, providing a more comprehensive and accurate basis for determining device status.

[0053] In one embodiment, the local trend consistency between the low-frequency sub-signal of the vibration signal and the velocity signal is obtained according to the following method: S321: Determine the change direction of the amplitude of the low-frequency sub-signal of the vibration signal and the velocity signal at each sampling point.

[0054] For the low-frequency sub-signal of the vibration signal : calculate The difference between the amplitude at each sample point and the amplitude at the previous sample point:

[0055] In the formula, for In the The amplitude at the sampling point and In the The amplitude at the sampling point The difference between.

[0056] like , In the The variation trend of the amplitude at each sampling point is increasing; like , In the The variation trend of the amplitude at each sampling point is decreasing; like , In the The variation trend of the amplitude at each sampling point remains unchanged.

[0057] For speed signal : calculate The difference between the amplitude at each sample point and the amplitude at the previous sample point:

[0058] In the formula, for In the The amplitude at the sampling point and In the The amplitude at the sampling point The difference between.

[0059] like , In the The variation trend of the amplitude at each sampling point is increasing; like , In the The variation trend of the amplitude at each sampling point is decreasing; like , In the The variation trend of the amplitude at each sampling point remains unchanged.

[0060] S322: Determine trend consistency of the amplitudes of the low-frequency sub-signal of the vibration signal and the velocity signal at the same sampling point according to the change direction of the amplitudes of the two at the same sampling point.

[0061] If the amplitudes of the low-frequency component of the vibration signal and the velocity signal at that sampling point change in the same direction, this indicates that the velocity signal and the low-frequency component of the vibration signal exhibit consistent trends at that sampling point. For example, vibration increases with increasing speed, or decreases with decreasing speed. This consistent trend conforms to the energy transfer laws under normal operating conditions. The macroscopic motion (speed) and low-frequency vibration of the mechanical system are driven by the same transmission chain, and instantaneous changes should maintain dynamic coordination without additional interference.

[0062] If the amplitudes of the low-frequency component of the vibration signal and the velocity signal at a given sampling point change in different directions, this indicates that the velocity signal and the low-frequency component of the vibration signal exhibit opposite trends at that sampling point. For example, vibration suddenly decreases when speed increases, or increases when speed decreases. This divergence violates the logic of normal energy transfer and usually indicates a localized transient anomaly, such as a sudden increase in speed but a sudden drop in vibration due to a stuck transmission component, or a sudden increase in vibration due to a sudden change in bearing clearance.

[0063] In one embodiment, the trend consistency of the amplitudes of the low-frequency sub-signal of the vibration signal and the velocity signal at the same sampling point is first calculated:

[0064] In the formula, for and In the The trend consistency of the amplitude at each sampling point, is a symbolic function.

[0065] when and When both are positive or negative at the same time, it means and In the The variation trend of the amplitude at each sampling point is the same. , ,at this time and In the The trend consistency of the amplitude at the sampling points is the greatest. .

[0066] when and One is positive and the other is negative, indicating and In the The changing trends of the amplitudes at the sampling points are different. , ,at this time and In the The trend consistency of the amplitude at the sampling points is the smallest. .

[0067] S323: Determine the local trend consistency of the low-frequency sub-signal of the vibration signal and the velocity signal according to the trend consistency of the amplitudes of the low-frequency sub-signal of the vibration signal and the velocity signal at the same sampling point.

[0068] In one embodiment, the local trend consistency between the low-frequency sub-signal of the vibration signal and the velocity signal is determined as follows: Calculate the mean trend consistency of the amplitude of the low-frequency sub-signal of the vibration signal and the velocity signal at all sampling points as the local trend consistency of the low-frequency sub-signal and the velocity signal. Specifically, obtain and speed signal At all sampling points , then calculate all The mean of and speed signal local trend consistency.

[0069] S33: Combine the global trend consistency and the local trend consistency to obtain the final trend consistency.

[0070] The global trend consistency and local trend consistency of the low-frequency sub-signal of the vibration signal and the velocity signal are fused to obtain the trend consistency of the low-frequency sub-signal and the velocity signal.

[0071] If the global trend consistency between the low-frequency sub-signal of the vibration signal and the speed signal is greater, and the local trend consistency is greater, then the low-frequency sub-signal of the vibration signal and the speed signal have consistent change trends both globally and locally, indicating that the high-speed frame stranding machine is more likely to be in a normal and stable operating state at the current moment. Conversely, if the global trend consistency between the low-frequency sub-signal of the vibration signal and the speed signal is less, and the local trend consistency is less, then the low-frequency sub-signal of the vibration signal and the speed signal have inconsistent change trends both globally and locally, indicating that the high-speed frame stranding machine is more likely to be in an abnormal operating state at the current moment.

[0072] Therefore, fusion can be achieved by multiplying the global trend consistency of the low-frequency sub-signal of the vibration signal and the speed signal, and the local trend consistency of the low-frequency sub-signal of the vibration signal and the speed signal, and the product is used as the trend consistency of the low-frequency sub-signal of the vibration signal and the speed signal.

[0073] for and For , the trend consistency is:

[0074] In this formula, for and The trend consistency, for and The local trend consistency, for and Global trend consistency.

[0075] In summary, through steps S31 to S33, by quantifying the correlation between the speed signal of the high-speed frame stranding machine and the low-frequency sub-signal of the vibration signal, it is possible to quickly determine whether the system is in a stable operating state from a macro-dynamic level, provide necessary pre-screening for the subsequent refined analysis of high-frequency fault characteristics, avoid redundant calculations of normal operating conditions, and accurately locate abnormal conditions that require in-depth evaluation.

[0076] S4: Determine the abnormality of the high-frequency sub-signal of the vibration signal through time-frequency analysis.

[0077] As mentioned earlier, the low-frequency sub-signal of the vibration signal primarily reflects the macroscopic motion characteristics of the equipment (such as periodic fluctuations in spool rotation and overall imbalance). Its trend consistency with the velocity signal can be used to determine whether macroscopic energy transfer is normal. However, early-stage equipment failures, such as minor impacts between bearing rollers and rails, localized wear on gear tooth surfaces, and slight looseness of connectors, often manifest themselves in high-frequency vibrations. These microscopic anomalies generate high-frequency impact or friction vibrations, which are far higher than the macroscopic motion frequencies covered by the low-frequency sub-signal and difficult to capture with conventional low-frequency velocity information. In the early stages of a failure, such anomalies have minimal impact on the macroscopic velocity signal and low-frequency vibrations, but they produce noticeable abnormal characteristics in the high-frequency vibrations.

[0078] Therefore, this step can keenly capture the early characteristics of equipment micro-faults by analyzing the abnormality of the high-frequency sub-signals of the vibration signal, filling the gap in macro trend analysis in identifying subtle anomalies.

[0079] In one embodiment, the method for obtaining the abnormality degree of the high frequency sub-signal of the vibration signal is as follows: Figure 3 As shown, including: S41: Perform time-frequency analysis on the high-frequency sub-signal.

[0080] The Fourier transform is a common time-frequency analysis technique. Its core purpose is to convert a time-domain signal (a signal that varies over time) into a frequency-domain representation, thereby revealing the signal's frequency components and their characteristics. The Fourier transform can be used to obtain the signal's frequency distribution and the energy at each frequency. This is a core method for analyzing a signal's frequency-domain characteristics in signal processing.

[0081] Therefore, Fourier transform is performed on each high-frequency sub-signal of the vibration signal to obtain all frequencies contained in each high-frequency sub-signal and the energy value at each frequency.

[0082] S42: Determine the signal impact strength according to the time domain characteristics of the high-frequency sub-signal.

[0083] When identifying abnormal characteristics of high-frequency components of vibration signals through time-frequency analysis, focusing on the quantification of impact intensity in the time domain is crucial. Mechanical faults (such as pitting impact between bearing rollers and rails, and meshing impact caused by cracks on gear teeth) are most notably characterized by non-stationary, transient impacts in high-frequency components. These impact signals are typically short-lived and exhibit significant amplitude fluctuations. Time-domain analysis can directly capture the nature of these transient fluctuations.

[0084] The peak value, the maximum amplitude of the signal in the time domain, directly corresponds to the energy release intensity of a single impact event and provides a direct reflection of the impact force. For mechanical failures, the peak value directly reflects the energy transfer from the fault source during the instantaneous impact, making it a key parameter for measuring the physical intensity of an impact.

[0085] Kurtosis is highly sensitive to extreme signal values. Under normal operating conditions, vibration signals typically exhibit a near-Gaussian distribution with a kurtosis close to 3. However, when a localized fault occurs, sparsely occurring shock signals thicken the tail of the probability distribution, significantly increasing the kurtosis. This characteristic effectively amplifies the difference between sparse shocks and the background steady vibration, accurately characterizing the degree to which shock events deviate from the norm. This overcomes the limitation that peak values ​​only reflect amplitude but fail to capture the frequency and distribution of shocks.

[0086] Therefore, for each high-frequency sub-signal: The deviation between the peak value and the reference peak value of the high-frequency sub-signal, as well as the deviation between the kurtosis of the high-frequency sub-signal and the reference kurtosis, are calculated. The product of these two deviations is used as the signal impact intensity of the high-frequency sub-signal. The intensity of the impact feature of the high-frequency sub-signal is quantified, which helps to sensitively identify the sparse impact signals caused by early mechanical failures.

[0087] For the vibration signal High-frequency sub-signals , the signal impact strength is:

[0088] In this formula, for The signal impact strength, for The peak value, The reference peak value is a vibration signal collected in advance under normal working conditions (the length is the same as the length of the vibration signal in step S1) as a reference vibration signal, and the peak value of each high-frequency sub-signal of the reference vibration signal is obtained according to the same operation as the present invention, and the average of the peak values ​​of all high-frequency sub-signals of the reference vibration signal is taken as the reference peak value. , for The kurtosis, The reference kurtosis is the mean of the kurtosis of all high-frequency sub-signals of the reference vibration signal.

[0089] In this formula, The vibration signal High-frequency sub-signals The deviation between the peak value and the reference peak value, The vibration signal High-frequency sub-signals The deviation between the kurtosis and the baseline kurtosis is amplified by multiplying the two together, improving the recognition of weak fault shocks and enhancing the sensitivity to impact faults.

[0090] The greater the deviation between the peak value of a high-frequency sub-signal of the vibration signal and the reference peak value, and the greater the deviation between the kurtosis of the high-frequency sub-signal and the reference kurtosis, the greater the signal impact strength of the high-frequency sub-signal, and the more likely it is that the impact characteristics of the signal are caused by a fault. Conversely, the smaller the signal impact strength of the high-frequency sub-signal, the more likely it is in a normal and stable operating state.

[0091] Then, we use the S-type function (such as sigmoid function) to Perform normalization operation (that is, and The product of is normalized), so that in Within the range.

[0092] S43: Determine the energy concentration level according to the frequency domain characteristics of the high-frequency sub-signal.

[0093] Frequency domain characteristics reflect the energy distribution of a signal across different frequency components. In mechanical fault diagnosis, the concentration of a signal's energy is a key indicator of the significance of the fault's characteristic frequency. Mechanical faults (such as bearing pitting and gear cracks) typically exhibit concentrated energy at specific frequencies.

[0094] Specifically, this step evaluates the energy concentration of the high-frequency sub-signal based on the entropy of the energy at all frequencies. The entropy of the energy at all frequencies of the high-frequency sub-signal is first normalized to obtain a normalized entropy value. The value obtained by subtracting the normalized entropy value from 1 is used as the energy concentration of the high-frequency sub-signal.

[0095] For the vibration signal High-frequency sub-signals , the entropy of its energy at all frequencies is:

[0096] In this formula, The vibration signal High-frequency sub-signals The entropy of the energy at all frequencies, is the sequence number of the frequency contained in the vibration signal, is the total number of frequencies contained in the vibration signal, for In the The energy value at the frequency, for The sum of the energy values ​​at all frequencies. Quantified The degree of disorder of the frequency domain energy distribution, The larger the value, the more dispersed the frequency domain energy distribution is. On the contrary, The smaller it is, the more concentrated the frequency domain energy distribution is.

[0097] Then, through right Perform normalization operation so that in Within the interval, is the total number of sampling points.

[0098] Finally, the energy concentration is:

[0099] In this formula, The vibration signal High-frequency sub-signals The energy concentration The vibration signal High-frequency sub-signals The entropy of the energy at all frequencies, is the total number of sampling points.

[0100] when The more concentrated the frequency domain energy distribution is, the greater the energy concentration is. The smaller, When the frequency domain energy distribution is completely concentrated at a single frequency, , at this time the energy concentration is the greatest, .when The more uniform the frequency domain energy distribution is, the smaller the energy concentration is. The bigger, when When the frequency domain energy distribution of is completely uniform, , at this time the energy concentration is the smallest, .

[0101] S44: Analyze the time-frequency consistency of high-frequency sub-signals.

[0102] Under normal operating conditions, high-frequency components of vibration signals typically exhibit stable time-frequency characteristics. However, faults (such as impact, friction, and cracks) can cause abnormal coupling between the time and frequency domain characteristics of high-frequency components. Mechanical faults (such as gear meshing shock) often produce responses in both the time domain (manifested as an impact pulse) and the frequency domain (manifested as a sudden increase in energy at a specific frequency component). If the time-domain impact characteristics of the high-frequency component coincide with the concentrated energy characteristics in the frequency domain, the likelihood of a fault is higher.

[0103] Therefore, the first coefficient of variation is calculated based on the amplitude of the high-frequency sub-signal at all sampling points, and the second coefficient of variation is calculated based on all frequencies of the high-frequency sub-signal. According to the difference between the first coefficient of variation and the second coefficient of variation, the difference is used as a negative correlation indicator of time-frequency consistency to quantify the time-frequency consistency.

[0104] For the vibration signal High-frequency sub-signals , its time-frequency consistency satisfies the following relationship:

[0105] In this formula, The vibration signal High-frequency sub-signals time-frequency consistency. for The first coefficient of variation of the vibration signal High-frequency sub-signals The coefficient of variation of all amplitudes is calculated from the amplitudes at all sampling points. The specific value is The ratio of the standard deviation of the amplitude at all sample points to the mean. for The second coefficient of variation of the vibration signal High-frequency sub-signals The coefficient of variation of all frequencies, The specific value is The ratio of the standard deviation of all frequencies to the mean, || is the absolute value symbol, is the natural exponential function.

[0106] In this formula, is the difference between the first coefficient of variation and the second coefficient of variation. The smaller the difference, the better the time-frequency consistency, and vice versa. Therefore, through the natural exponential function Build and negative correlation.

[0107] In short, Essentially, it is quantitative High-frequency sub-signals If abnormal fluctuations appear in both the time domain and the frequency domain at the same time, the changes in the two are considered to be consistent. This abnormal fluctuation is more likely to be a common response caused by the fault).

[0108] S45: Fusion signal impact strength, energy concentration and time-frequency consistency to determine the degree of abnormality.

[0109] From the theoretical perspective of signal processing, the signal impact intensity reflects the transient impact component of the signal, which is directly related to the pulse response caused by local faults (such as bearing pitting). The energy concentration degree can effectively identify the resonance frequency band offset or abnormal energy aggregation phenomenon. The time-frequency consistency quantifies the stability of the time-frequency distribution and can capture the modulation characteristics under non-stationary working conditions. The fusion of these three indicators to determine the degree of abnormality is consistent with the physical mechanism of the fault and improves the detection robustness through feature complementarity.

[0110] Specifically, for each high-frequency sub-signal of the vibration signal, the product of the signal impact intensity, energy concentration and time-frequency consistency of the high-frequency sub-signal is used as the abnormality degree of the high-frequency sub-signal.

[0111] The vibration signal High-frequency sub-signals The degree of abnormality is:

[0112] In this formula, The vibration signal High-frequency sub-signals The degree of abnormality, for The energy concentration for The time-frequency consistency of for signal impact strength.

[0113] Finally, the average of the abnormality levels of all high-frequency sub-signals is taken as the abnormality level of the high-frequency sub-signal of the vibration signal, that is, the average of the abnormality levels of all high-frequency sub-signals of the vibration signal is taken as the abnormality level of the high-frequency sub-signal of the vibration signal. The mean of the high-frequency sub-signal of the vibration signal is used as the abnormality degree of the high-frequency sub-signal of the vibration signal.

[0114] This is because the anomaly of a single high-frequency sub-signal may be local interference, and the mean can integrate information from multiple frequency bands. If multiple sub-signals show anomalies, the mean will increase significantly, better reflecting the overall high-frequency anomaly trend of the device. If only individual high-frequency sub-signals are abnormal, the mean will be diluted, reducing local interference.

[0115] S5: Based on the trend consistency of the low-frequency sub-signal and the speed signal, as well as the abnormality degree of the high-frequency sub-signal, the operating condition abnormality index of the high-speed frame stranding machine is comprehensively determined.

[0116] Low-frequency information (low-frequency sub-signals of the speed and vibration signals) during high-speed stranding machine operation reflects the machine's macroscopic operating state, specifically the degree of matching between speed and low-frequency vibration. A higher degree of matching between speed and low-frequency vibration indicates more coordinated changes between the two, more consistent with normal load variations. This translates to smoother and more normal operation. A lower degree of matching between speed and low-frequency vibration indicates less coordinated changes between the two, less consistent with normal load variations. This indicates a potential mechanical failure, such as loose transmission system, bearing wear, or sudden load changes. This indicates a significant abnormality requiring a priority warning. High-frequency information (high-frequency sub-signals of the vibration signal) during high-speed stranding machine operation reflects microscopic transient anomalies, such as impact, friction, and localized damage, and is typically used for early warning.

[0117] Therefore, if the trend consistency between the low-frequency sub-signal of the vibration signal and the speed signal is better, and the degree of abnormality of the high-frequency sub-signal of the vibration signal is smaller, the high-speed frame stranding machine is more likely to be in a normal and stable operating state. If the trend consistency between the low-frequency sub-signal of the vibration signal and the speed signal is worse, and the degree of abnormality of the high-frequency sub-signal of the vibration signal is greater, it means that the high-speed frame stranding machine is more likely to have a fault.

[0118] Specifically, the abnormal operating index of the high-speed frame stranding machine is determined as follows: The trend consistency of the low-frequency sub-signal and the speed signal is used as the first parameter, the abnormality degree of the high-frequency sub-signal is used as the second parameter, the first parameter is converted into a negative correlation index of the operating condition abnormality index through a subtraction operation, and the product of the second parameter and the negative correlation index after performing a plus operation is used as the operating condition abnormality index.

[0119] It can be expressed as:

[0120] Where, For high-speed frame stranding machine The abnormal working condition index at the moment, Low-frequency sub-signal and speed signal Trend consistency (first parameter), It is the abnormality degree of the high-frequency sub-signal of the vibration signal (the second parameter).

[0121] In this formula, through Build and There is a negative correlation between The larger it is, the closer it is to 1, indicating a low-frequency sub-signal and speed signal The stronger the trend consistency, the more obvious the abnormality of the high-speed frame stranding machine. The smaller it is, the closer it is to 0, and the abnormal working condition index is Take the lead, at this time The larger the value is, the greater the abnormality of the high-frequency sub-signal of the vibration signal is, and the more likely it is that the high-speed frame stranding machine has a microscopic fault. The smaller it is, the closer it is to 0, indicating a low-frequency sub-signal and speed signal The weaker the trend consistency, the more obvious abnormality in the working condition of the high-speed frame stranding machine exists. The larger the value, the greater the abnormal working condition index. Take the lead, if The larger the value, the further the macro abnormality is amplified, and the sensitivity of fault warning is enhanced. The smaller, the avoid Macroscopic anomalies The unreasonable reduction of macro abnormalities ensures that fault warnings are triggered first.

[0122] In short, this design can not only capture obvious anomalies when the low-frequency trends are inconsistent, but also discover potential faults through high-frequency information when the low-frequency trends are consistent. It realizes the effective integration of two-dimensional anomaly monitoring and can balance the timeliness and accuracy of fault detection.

[0123] S6: Control the high-speed frame stranding machine according to the abnormal working condition index.

[0124] The first and second thresholds for the operating condition anomaly index are set to 0.3 and 0.7, respectively, both of which are empirical values. At the current moment, if the operating condition anomaly index obtained from steps S1 to S5 is less than 0.3, the operating condition is determined to have no early-stage anomaly, and the high-speed frame stranding machine continues to operate with the parameters corresponding to the current operating condition. If the operating condition anomaly index obtained from steps S1 to S5 is greater than 0.3 and less than or equal to 0.7, the operating condition is determined to have a minor early-stage fault, and the parameters corresponding to the current operating condition are reduced by a preset percentage of 10% (empirical value), and operation continues, with a prompt to personnel to continue monitoring. If the operating condition anomaly index obtained from steps S1 to S5 is greater than 0.7, the high-speed frame stranding machine is determined to have a significant early-stage fault, and the machine is immediately reduced to idle speed, with a prompt to personnel to stop and inspect.

[0125] An intelligent online control system for a high-speed frame stranding machine of the present invention includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the operations of steps S1 to S6, thereby performing intelligent online control of the high-speed frame stranding machine.

[0126] While various embodiments of the present invention have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only.

Claims

1. An intelligent online control method for a high-speed frame stranding machine, characterized in that: include: Collect the speed signal and vibration signal of the high-speed frame stranding machine during operation, and obtain the high-frequency and low-frequency sub-signals of the vibration signal through signal decomposition technology; For the low-frequency sub-signal, the trend consistency between the low-frequency sub-signal and the speed signal is determined through trend analysis; For high-frequency sub-signals, perform time-frequency analysis, including: In the time domain, the signal impact strength of the high-frequency sub-signal is determined by the peak value and kurtosis of the high-frequency sub-signal, and a first coefficient of variation is calculated based on the amplitude of the high-frequency sub-signal at all sampling points; in the frequency domain, the energy concentration of the high-frequency sub-signal is evaluated based on the entropy value of the energy of the high-frequency sub-signal at all frequencies, and a second coefficient of variation is calculated based on all frequencies of the high-frequency sub-signal; and the time-frequency consistency of the high-frequency sub-signal is determined based on the difference between the first coefficient of variation and the second coefficient of variation; By fusing the signal's impact strength, energy concentration, and time-frequency consistency, the abnormality of the high-frequency sub-signal is determined; According to the trend consistency of the low-frequency sub-signal and the speed signal, as well as the abnormality degree of the high-frequency sub-signal, the abnormal working condition index of the high-speed frame stranding machine is comprehensively determined, and the high-speed frame stranding machine is controlled according to the abnormal working condition index.

2. The intelligent online control method according to claim 1, characterized in that: The trend consistency of the low-frequency sub-signal and the speed signal is determined based on the following method: Based on the amplitudes of the low-frequency sub-signal and the speed signal at all sampling points, the Pearson correlation coefficient of the low-frequency sub-signal and the speed signal is calculated, and the Pearson correlation coefficient is used as the global trend consistency; Determining the change direction of the amplitude of the low-frequency sub-signal and the velocity signal at each sampling point based on the difference between the amplitude of the low-frequency sub-signal and the velocity signal at each sampling point and the amplitude at the previous sampling point; Determining the local trend consistency of the amplitudes of the low-frequency sub-signal and the velocity signal according to the change direction of the amplitudes of the low-frequency sub-signal and the velocity signal at the same sampling point; The global trend consistency and the local trend consistency index are fused to obtain the trend consistency of the low-frequency sub-signal and the speed signal.

3. The intelligent online control method according to claim 1, characterized in that: The signal impact strength of the high-frequency sub-signal is determined based on the following method: A first deviation between the peak value of the high-frequency sub-signal and the reference peak value, and a second deviation between the kurtosis of the high-frequency sub-signal and the reference kurtosis are calculated, and a normalized value of the product of the first deviation and the second deviation is used as the signal impact strength of the high-frequency sub-signal.

4. The intelligent online control method according to claim 1, characterized in that: The first coefficient of variation is the ratio of the standard deviation of the amplitude of the high-frequency sub-signal at all sampling points to the mean, and the second coefficient of variation is the ratio of the standard deviation of all frequencies of the high-frequency sub-signal to the mean.

5. The intelligent online control method according to claim 1, characterized in that: The time-frequency consistency of the high-frequency sub-signal is determined based on the following method: The difference between the first coefficient of variation and the second coefficient of variation was calculated and used as a negative correlation indicator of time-frequency consistency to quantify time-frequency consistency.

6. The intelligent online control method according to claim 1, characterized in that: The energy concentration of the high-frequency sub-signal is determined based on the following method: The entropy value of the energy of the high-frequency sub-signal at all frequencies is normalized to obtain a normalized entropy value, and the value obtained by subtracting the normalized entropy value from 1 is used as the energy concentration degree of the high-frequency sub-signal.

7. The intelligent online control method according to claim 1, characterized in that: The degree of abnormality of the high-frequency sub-signal is determined based on the following method: The product of signal impact intensity, energy concentration and time-frequency consistency is taken as the abnormality degree of high-frequency sub-signal.

8. The intelligent online control method according to claim 1, characterized in that: The abnormal operating index of the high-speed frame stranding machine is determined based on the following method: The trend consistency of the low-frequency sub-signal and the speed signal is used as the first parameter, the abnormality degree of the high-frequency sub-signal is used as the second parameter, the first parameter is converted into a negative correlation index of the operating condition abnormality index through a subtraction operation, and the product of the second parameter and the negative correlation index after performing a plus operation is used as the operating condition abnormality index.

9. The intelligent online control method according to claim 1, characterized in that: The method for controlling the high-speed frame stranding machine is: Set the first and second thresholds of the abnormal working condition index; If the operating condition abnormality index is less than the first-level threshold, it is determined that no early abnormality has occurred in the operating condition, and the high-speed frame stranding machine continues to operate while maintaining the parameters corresponding to the current operating condition; If the operating condition abnormality index is greater than the first-level threshold and not greater than the second-level threshold, it is determined that the operating condition has a minor early fault, and the parameters corresponding to the current operating condition are reduced by a preset percentage to continue operation, and the staff is prompted to continue paying attention; If the operating condition abnormality index is greater than the secondary threshold, it is determined that the operating condition has an obvious early fault, the speed is immediately reduced to idle speed, and the staff is prompted to stop and check.

10. An intelligent online control system for a high-speed frame stranding machine, characterized in that: The intelligent online control system includes a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of the online control method according to any one of claims 1 to 9.

Images