Method and system for real-time control of knitting machine motion stability

Abstract

The present application relates to the technical field of control of circular knitting machines, and in particular to a real-time control method and system for the motion stability of a circular knitting machine. The present application extracts the fluctuation feature vectors of the vibration data in the vibration period and divides the normal period and the abnormal period by means of the similarity between the fluctuation feature vectors. Further, the vibration stability coefficients of each monitoring point are obtained according to the distribution of the normal period and the abnormal period and in combination with the similarity of the fluctuation feature vectors. Further, the overall stability coefficient of the needle cylinder is obtained according to the fluctuation similarity between the vibration data, in combination with the distribution similarity of the normal period and all the vibration stability coefficients. Further, the hardware abnormal factor is obtained according to the fluctuation intensity feature of the voltage data and in combination with the overall stability coefficient. Finally, the control mode of the circular knitting machine is switched based on the hardware abnormal factor, the hardware abnormality is more accurately and comprehensively identified, the control and maintenance are timely switched, and the operation stability and production quality are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of knitting circular machine control, and particularly relates to a real-time control method and system for knitting circular machine motion stability. BACKGROUND

[0002] The knitting circular machine is a kind of knitting equipment for producing weft-knitted fabric, and becomes the core equipment in the weft-knitted field of clothing textiles, home textiles, industrial textiles, etc. due to its high production capacity and diverse fabric adaptability.

[0003] During the long-term operation of mechanical components, collisions and friction anomalies occur between each other due to continuous friction, vibration and self-wear and aging, and this wear and aging is a gradual accumulation process. It is difficult to accurately predict when serious collision and friction anomaly problems will occur through simple monitoring means, and it is difficult to prevent. It is difficult to balance maintenance resource waste and maintenance timeliness with fixed maintenance period. The hardware wear and aging has a greater impact on the overall structure and motion state, affecting the production stability and production quality of the knitting circular machine. SUMMARY

[0004] In order to solve the technical problems that the hardware wear of the knitting circular machine is difficult to predict and timely maintained, and affects the production stability and production quality of the knitting circular machine, the purpose of the present application is to provide a real-time control method and system for knitting circular machine motion stability, and the technical solution adopted is as follows:

[0005] A real-time control method for knitting circular machine motion stability, the method comprises:

[0006] Obtain vibration data and voltage data of each monitoring point of the knitting circular machine; in the current preset historical neighborhood, according to the frequency domain distribution of the vibration data, obtain the vibration period of each monitoring point;

[0007] Extract the fluctuation feature vector of the vibration data in the vibration period; for each monitoring point, according to the similarity between the fluctuation feature vectors of different vibration periods, divide the vibration period into normal period and abnormal period; according to the distribution of the normal period and the abnormal period, combined with the similarity features of the fluctuation feature vectors between the abnormal periods, obtain the vibration stability coefficient of each monitoring point in the current;

[0008] According to the fluctuation similarity of the vibration data between the monitoring points, combined with the distribution similarity of the normal period and all the vibration stability coefficients, obtain the overall stability coefficient of the needle cylinder; according to the fluctuation intensity feature of the voltage data in the preset historical neighborhood, combined with the overall stability coefficient, obtain the hardware abnormal factor;

[0009] Switch a control mode of the circular knitting machine based on the hardware exception factor.

[0010] Further, the method for obtaining the fluctuation feature vector comprises:

[0011] The maximum amplitude of the vibration data in each period and the time interval between the first trough and the first peak form a fluctuation feature vector.

[0012] Further, the method for dividing the vibration period into normal periods and abnormal periods comprises:

[0013] At the same monitoring point, the average of the Euclidean distances between the fluctuation feature vector of each vibration period and the fluctuation feature vectors of all other vibration periods is taken as the vector difference coefficient of each vibration period. The vibration period with a vector difference coefficient less than a preset vector difference threshold is marked as a normal period, and the remaining vibration periods are marked as abnormal periods.

[0014] Further, the method for obtaining the vibration stability coefficient comprises:

[0015] At the same monitoring point, the fluctuation feature vectors of the abnormal periods are clustered. In each cluster, an important influence factor of each cluster is obtained according to the distribution of the abnormal periods.

[0016] According to the number of normal periods, the time domain proportion of the maximum continuous duration of the normal periods, and in combination with all the important influence factors, a vibration stability coefficient of the monitoring point is obtained.

[0017] Further, the method for obtaining the important influence factor comprises:

[0018] In each cluster, an important influence factor of each cluster is obtained according to the minimum time interval of the time-sequentially adjacent abnormal periods, in combination with the maximum time interval of the first and last abnormal periods in the time domain and the number of abnormal periods in the cluster.

[0019] Further, the method for obtaining the overall stability coefficient comprises:

[0020] The Pearson correlation coefficient of the vibration data of any two monitoring points is obtained.

[0021] The proportion of the number of normal periods of any two monitoring points in the time domain is taken as a similarity correction coefficient of the corresponding two monitoring points.

[0022] The Pearson correlation coefficients and the similarity correction coefficients of each monitoring point and all other monitoring points are fused, and in combination with all the vibration stability coefficients, an overall stability coefficient of the needle cylinder is obtained.

[0023] Further, the method for obtaining the hardware abnormality factor comprises:

[0024] In a preset neighborhood of each voltage extreme point, a local instability coefficient of each voltage extreme point is obtained according to a difference of voltage amplitudes adjacent in time domain; and an unstable voltage point is screened out based on the local instability coefficient;

[0025] The number of the unstable voltage point and the overall stability coefficient are fused to obtain the hardware abnormality factor.

[0026] Further, the method for obtaining the vibration period comprises:

[0027] The vibration data of each monitoring point in a current preset historical neighborhood is subjected to Fourier transform to obtain a dominant frequency with the highest amplitude in a frequency spectrum, a period length is obtained based on the dominant frequency, the preset historical neighborhood is divided by the period length corresponding to each monitoring point, and a vibration period of each monitoring point is obtained.

[0028] Further, the method for switching a control mode of the circular knitting machine based on the hardware abnormality factor comprises:

[0029] When the hardware abnormality factor is greater than a preset abnormality threshold, the circular knitting machine is controlled by fuzzy control.

[0030] The application further provides a real-time control system for motion stability of a circular knitting machine, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of any one of the real-time control methods for motion stability of a circular knitting machine.

[0031] The application has the following beneficial effects:

[0032] The application firstly acquires vibration data and voltage data and acquires the vibration period of each monitoring point according to the frequency domain characteristics of the vibration data, providing an analysis basis for the subsequent; further extracts the fluctuation feature vector of the vibration data in the vibration period, divides the vibration period into normal period and abnormal period according to the similarity between the fluctuation feature vectors of different vibration periods, and helps to accurately extract the real vibration characteristics of the circular knitting machine; further acquires the current vibration stability coefficient of each monitoring point according to the distribution of the normal period and the abnormal period, and combines the similarity features of the fluctuation feature vectors between the abnormal periods, which represents the vibration stability degree and provides a basis for subsequent calculation of the overall stability coefficient; further acquires the overall stability coefficient of the needle cylinder according to the fluctuation similarity of the vibration data between the monitoring points, combines the distribution similarity of the normal period and all vibration stability coefficients, avoids the excessive influence of single monitoring point anomaly on the overall judgment, and can reveal the damage degree of local anomaly to the overall motion stability from the global level; further acquires the hardware anomaly factor according to the fluctuation intensity features of the voltage data in the preset historical neighborhood, combines the overall stability coefficient, avoids misjudging the external voltage mutation as mechanical failure, and makes the hardware anomaly factor truly reflect the health status of the overall hardware of the circular knitting machine; finally, the control mode of the circular knitting machine is switched based on the hardware anomaly factor. The application analyzes the vibration data of each monitoring point, acquires the vibration period, stability coefficient and overall stability coefficient, combines the voltage fluctuation features to obtain the hardware anomaly factor, more accurately and comprehensively identifies the hardware anomaly, switches the control and maintenance in time, and improves the operation stability and production quality. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0034] Figure 1 A flow chart of a real-time control method for motion stability of a circular knitting machine provided by an embodiment of the present application;

[0035] Figure 2 A flow chart of a vibration stability coefficient acquisition method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined object, the specific implementation, structure, features and effects of the real-time control method and system for the motion stability of a circular knitting machine according to the present application are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0037] Unless otherwise defined, 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 application belongs.

[0038] The specific scheme of the real-time control method and system for the motion stability of a circular knitting machine provided by the present application is described in detail below in combination with the drawings.

[0039] Please refer to Figure 1 which shows a flowchart of a real-time control method for the motion stability of a circular knitting machine according to an embodiment of the present application, which specifically includes:

[0040] Step S1: Obtain the vibration data and voltage data of each monitoring point of the circular knitting machine; in the current preset historical neighborhood, obtain the vibration period of each monitoring point according to the frequency domain distribution of the vibration data.

[0041] In order to control the motion stability of the circular knitting machine in real time, it is necessary to first determine the vibration data and voltage data of the key parts of the circular knitting machine.

[0042] In an embodiment of the present application, considering that the needle cylinder is a key component for the circular knitting machine to realize the knitting function, the stable operation of the needle cylinder is crucial for the control of the yarn tension, and the vibration of the needle cylinder directly reflects the stability of the mechanical structure of the circular knitting machine. During the operation of the needle cylinder, its vibration is affected by multiple mechanical components, such as the matching accuracy of the needle cylinder and the cam seat, the positional relationship between the needle cylinder and the roller, and the manufacturing accuracy and installation quality of the needle cylinder itself. If these mechanical components have problems such as wear, looseness or installation deviation, it will first be reflected in the vibration of the needle cylinder, so the needle cylinder is selected as the main monitoring part of the circular knitting machine.

[0043] Sensor selection and installation: piezoelectric acceleration vibration sensor is selected. Such sensors have the characteristics of high sensitivity and wide frequency response range, and can accurately perceive small and rapidly changing vibration signals, meeting the measurement needs of the complex vibration environment of the circular knitting machine. For example, a piezoelectric acceleration vibration sensor with a measurement range of ±50g, a sensitivity of 100mV / g, and a frequency response range of 0.5Hz-10kHz can be selected to ensure that the vibration information of the key parts of the circular knitting machine under different operating conditions can be captured; the sensor is installed at 3-5 evenly selected installation points on the outer wall of the needle cylinder in the circumferential direction, and the sensor is firmly installed using a dedicated magnetic base or strong glue;

[0044] A voltage sensor is selected and installed at the power supply inlet of the circular knitting machine to obtain the voltage signal directly supplied to the machine, which truly reflects the actual working voltage of the machine. The sampling frequency can be the same as the frequency setting of the vibration sensor.

[0045] Data acquisition and storage: the interval of vibration data acquisition is set to 10ms to ensure that sufficient effective vibration data can be obtained; the analog signals output by each vibration sensor are collected in real time and converted into digital signals, which are transmitted to the computer together with the voltage signal; the vibration data and voltage data of each monitoring point and the vibration curve are obtained.

[0046] It should be noted that the collected data is continuous running data, and in other embodiments of the present application, the implementer can adjust the data collection frequency, sensor selection and installation as needed, which will not be described again.

[0047] Since the needle cylinder rotates in a circular motion, each part is uniformly stressed during rotation, and the vibration curve presents a periodic characteristic closely related to the rotation period of the needle cylinder. In a complete rotation period, the shape of the vibration curve is relatively fixed and periodically repeated, so the vibration period of the vibration data is first extracted. Considering that the frequency domain of the vibration data contains the frequency characteristics of the vibration data, in the current preset historical neighborhood, the frequency domain distribution of the vibration data is considered to avoid the deviation of period recognition caused by instantaneous fluctuations or single-point interference, and the vibration period of each monitoring point is obtained to provide a basis for subsequent analysis.

[0048] Preferably, in one embodiment of the present application, the preset historical neighborhood is the historical adjacent 10 minutes before the current time;

[0049] The vibration data of each monitoring point in the current preset historical neighborhood is subjected to Fourier transform, the frequency with the highest amplitude in the frequency spectrum is obtained as the dominant frequency, and the period length is obtained based on the dominant frequency. The preset historical neighborhood is divided according to the period length corresponding to each monitoring point, and the vibration period of each monitoring point is obtained.

[0050] The reciprocal of the dominant frequency is taken as the cycle length, the preset historical neighborhood of the corresponding monitoring point is divided by the cycle length starting from the first time of the preset historical neighborhood, and the vibration cycle of each monitoring point is obtained.

[0051] It should be noted that the Fourier transform is prior art, and the analysis process of obtaining the hardware abnormal factor and determining the switching control mode of the circular knitting machine is consistent each time. Only one example is described herein. The determination frequency can be set to 10 minutes. In other embodiments of the present application, the implementer can adjust the length of the preset historical neighborhood and the analysis frequency of the hardware abnormal factor.

[0052] Step S2: Extract the fluctuation feature vector of the vibration data in the vibration cycle; for each monitoring point, according to the similarity between the fluctuation feature vectors of different vibration cycles, the vibration cycle is divided into normal cycle and abnormal cycle; according to the distribution of normal cycle and abnormal cycle, combined with the similarity of the fluctuation feature vectors between the abnormal cycles, the vibration stability coefficient of each monitoring point in the current is obtained.

[0053] During the operation of the machine, some non-periodic vibrations will be caused by factors such as yarn tension fluctuation and uneven needle wear, so it is necessary to distinguish all the vibration cycles in the vibration data. First, the fluctuation feature vector of the vibration data in the vibration cycle is extracted to provide a basis for dividing different cycles.

[0054] Preferably, in one embodiment of the present application, the maximum amplitude of the vibration data reflects the peak intensity in the cycle, which can intuitively reflect the maximum impact or energy concentration of the mechanical parts in the cycle; and the time interval between the first trough and the first peak reflects the response speed and frequency characteristics of the vibration signal at the beginning of the cycle, which can sensitively capture the dynamic changes caused by part wear, looseness or external disturbance.

[0055] Therefore, the amplitude corresponding to the maximum value of the vibration data in each cycle, and the time interval between the first trough and the first peak constitute the fluctuation feature vector, which provides a low-dimensional, easy-to-handle but information-rich feature representation for subsequent similarity measurement and stability coefficient calculation.

[0056] It is considered that in the operation of the machine, the vibration under ideal state presents stable and repeated periodicity, and the fluctuation feature vectors of different cycles should have high similarity; therefore, for each monitoring point, according to the similarity between the fluctuation feature vectors of different vibration cycles, the periodic stable state and the non-periodic abnormal state can be effectively distinguished, the vibration cycle is divided into normal cycle and abnormal cycle, which helps to accurately extract the real vibration characteristics of the circular knitting machine and reduces misjudgment and omission.

[0057] Preferably, in one embodiment of the present application, the Euclidean distance can be used to measure the difference between vectors, which reflects the similarity between fluctuation feature vectors from the side, so under the same monitoring point, the average value of the Euclidean distance between the fluctuation feature vector of each vibration period and the fluctuation feature vectors of all other vibration periods is taken as the vector difference coefficient of each vibration period; the vibration period with a vector difference coefficient less than a preset vector difference threshold is marked as a normal period, and the remaining vibration periods are marked as abnormal periods.

[0058] As an example, the preset vector difference threshold is 0.7, the larger the vector difference coefficient, the greater the Euclidean distance corresponding to a certain vibration period and all other vibration periods, the greater the difference, and the more likely it is an abnormal period, so the vector difference coefficient is linearly normalized in the corresponding data dimension, compared with the preset vector difference threshold, to divide the normal period and the abnormal period.

[0059] Among them, in order to prevent the influence of different data magnitudes in two dimensions, the maximum amplitude of the vibration data in each period and the time interval between the first trough and the first peak can be linearly normalized in the corresponding data dimension before constructing the fluctuation feature vector.

[0060] It should be noted that the data dimension used for normalization is the data dimension composed of all the same type of feature parameters of all monitoring points, for example, when normalizing the vector difference coefficient, the data dimension composed of the vector difference coefficients of all vibration periods of all monitoring points is used. Since linear normalization only needs the maximum value, the minimum value and the value to be normalized in the data dimension, the required storage and computing resources are less, and other normalization methods in the embodiments of the present application can use this method, which is a well-known technology and will not be described in detail.

[0061] In another embodiment of the present application, considering that the cosine similarity can also be used to measure the similarity between vectors, the cosine similarity can be used to replace the Euclidean distance, the average value of the cosine similarity between the fluctuation feature vector of each vibration period and the fluctuation feature vectors of all other vibration periods is taken as the vector similarity coefficient of each vibration period; the vibration period with a vector difference coefficient greater than a preset vector similarity threshold is marked as a normal period, and the remaining vibration periods are marked as abnormal periods.

[0062] Among them, the preset vector similarity threshold can be 0.5, and in other embodiments of the present application, the implementer can adjust the preset vector similarity threshold and the preset vector difference threshold.

[0063] The distribution of normal periods can reflect the stability of the monitoring point during operation, and the distribution of abnormal periods can reflect the concentration and evolution trend of abnormal periods, and reflect potential vibration instability factors. Therefore, according to the distribution of normal periods and abnormal periods, and in combination with the similar features of the fluctuation feature vectors between abnormal periods, the vibration stability coefficient of each monitoring point at the current time is obtained, which represents the vibration stability degree, and provides accurate and reliable input for subsequent calculation of the overall stability coefficient, thereby improving the reliability and accuracy of real-time control of the circular knitting machine.

[0064] Preferably, in one embodiment of the present application, please refer to Figure 2 which shows a flow chart of a vibration stability coefficient acquisition method provided by one embodiment of the present application, and specifically includes:

[0065] Step S201: Clustering the fluctuation feature vectors of abnormal periods under the same monitoring point; and in each cluster, according to the distribution of abnormal periods, obtaining the important influence factor of each cluster.

[0066] For all abnormal periods, the reasons for their formation are not exactly the same, and it is possible that sudden loosening of connecting parts, sudden knotting or foreign matter of yarn, instantaneous voltage mutation, etc. will cause abnormal vibration of the needle cylinder, and the characteristics of abnormal vibration caused by different types of reasons are also different. Therefore, the fluctuation feature vectors of abnormal periods are clustered to show the similar features of the fluctuation feature vectors between abnormal periods in a clustered manner. The fluctuation feature vectors of abnormal periods in each cluster are close to consistent, and it can be considered that the abnormal vibration causes in one cluster are the same.

[0067] It is considered that the smaller the minimum time interval of time-series adjacent abnormal periods in a cluster is, the more continuous and concentrated the abnormal periods are. The larger the maximum time interval of the first and last abnormal periods in the time domain is, the longer the occurrence time is. At the same time, the more the number of abnormal periods in the cluster is, the more important the reason for this type of abnormality is.

[0068] Based on this, in each cluster, according to the minimum time interval of time-series adjacent abnormal periods, in combination with the maximum time interval of the first and last abnormal periods in the time domain, and the number of abnormal periods in the cluster, the important influence factor of each cluster is obtained.

[0069] As an example, in each cluster, the abnormal periods are sorted in time sequence. For two time-series adjacent abnormal periods, the time interval between the time domain end of the time-series minimum abnormal period and the time domain start of the time-series maximum abnormal period is taken as the minimum time interval. The time interval between the time domain start of the time-series minimum abnormal period in the cluster and the time domain end of the time-series maximum abnormal period is taken as the maximum time interval.

[0070] The reciprocal of the average of the minimum time interval of all time-sequentially adjacent abnormal cycles in each cluster, the maximum time interval and the number of abnormal cycles in the cluster are multiplied as the important influence factor of each cluster.

[0071] Wherein, the two-dimensional fluctuation feature vector is taken as the two-dimensional coordinates of the abnormal cycle, the elbow method and the K-means clustering algorithm are adopted to cluster the fluctuation feature vectors of the abnormal cycles; the reciprocal of the average of the minimum time interval and the maximum time interval show the distribution of the abnormal cycles.

[0072] It should be noted that the elbow method and the K-means clustering algorithm are prior art, and in other embodiments of the present application, the implementer can also use other existing clustering algorithms such as DBSCAN, which will not be described again.

[0073] Step S202: According to the number of normal cycles, the time domain proportion of the maximum continuous length of normal cycles, and in combination with all important influence factors, the vibration stability coefficient of the monitoring point is obtained.

[0074] Considering that the larger the number of normal cycles in the preset historical neighborhood of the monitoring point is, the longer the overall duration of normal vibration is, and the more stable the vibration is; at the same time, the larger the time domain proportion of the maximum continuous length of normal cycles is, the longer the maximum duration of normal vibration is, and the more stable the vibration is; and the smaller the important influence factor is, the smaller the influence of abnormal cycles is, and the more stable the vibration is, so the vibration stability coefficient of the monitoring point is obtained by fusing the three.

[0075] As an example, when there is no abnormal cycle between the adjacent two normal cycles, it is considered that the adjacent two normal cycles are continuous, thereby determining the maximum continuous length of all normal cycles, determining the maximum continuous length of the normal cycle with the maximum continuous length, and taking the proportion of the maximum continuous length occupying the preset historical neighborhood as the time domain proportion of the maximum continuous length of the normal cycle;

[0076] For each monitoring point, the product of the number of normal cycles and the time domain proportion of the maximum continuous length of normal cycles is taken as the numerator, the average of the important influence factors of all clusters is taken as the denominator, and the fractional ratio is taken as the vibration stability coefficient of the monitoring point.

[0077] Wherein, the number of normal cycles and the time domain proportion of the maximum continuous length of normal cycles show the distribution of normal cycles.

[0078] Step S3: According to the fluctuation similarity of the vibration data between the monitoring points, in combination with the distribution similarity of the normal cycles and all vibration stability coefficients, the overall stability coefficient of the needle cylinder is obtained; according to the fluctuation intensity feature of the voltage data in the preset historical neighborhood, in combination with the overall stability coefficient, the hardware abnormal factor is obtained.

[0079] A plurality of vibration sensors are evenly arranged on the outer wall of the needle cylinder in the circumferential direction. The needle cylinder rotates as a whole in the circumferential direction, and each part rotates synchronously around the central axis. The driving force and the constraint condition are similar. When the circular knitting machine is running normally, the vibration rules of different positions of the needle cylinder are generally consistent. When the circular knitting machine is not running stably, the vibration rules of different positions of the needle cylinder are different, and the interference of mechanical parts is intensified.

[0080] For example, the relative positions of the components are more likely to change, and scraping may occur at a gap that is too small, causing the local force of the needle cylinder to change sharply, resulting in abnormal vibration at the point, which is completely different from the vibration rules of other points, and further affecting the movement stability of the entire machine. Or the needle cylinder and other components (such as sinkers and yarn guides) around it may have positional deviation due to unstable movement of the machine, and are more likely to have intermittent collisions. Each collision on a specific point of the needle cylinder will cause a sudden change in the vibration of the point.

[0081] Therefore, according to the similarity of the fluctuation of the vibration data between the monitoring points, in combination with the similarity of the distribution of the normal period and all vibration stability coefficients, the overall stability coefficient of the needle cylinder is obtained, the excessive influence of a single monitoring point anomaly on the overall judgment is avoided, the damage degree of the local anomaly to the overall movement stability is revealed from the global level, and a foundation is laid for subsequent acquisition of hardware anomaly factors and switching of control modes.

[0082] Preferably, in an embodiment of the present application, considering that the Pearson correlation coefficient can measure the fluctuation similarity of the vibration data between different monitoring points, the closer the Pearson correlation coefficient is to 1, the greater the fluctuation similarity of the vibration data is. Therefore, the Pearson correlation coefficient of the vibration data of any two monitoring points is obtained. At the same time, in order to prevent the Pearson correlation coefficient from being negative due to extreme cases and affecting the subsequent logical relationship, the Pearson correlation coefficient is taken as the independent variable and is mapped through the ReLU (x) function, ReLU (x) = Max (0, x);

[0083] Considering that the more normal periods that coincide in time between the monitoring points, the more similar the vibration characteristics of the two are, the number of normal periods that coincide in time between any two monitoring points is taken as a similarity correction coefficient of the two monitoring points, representing the distribution similarity of the normal period;

[0084] Further, the Pearson correlation coefficient and the similarity correction coefficient of each monitoring point and all other monitoring points are fused, and the overall stability coefficient of the needle cylinder is obtained in combination with all vibration stability coefficients.

[0085] As an example, the average value of the product of the Pearson correlation coefficient and the similarity correction coefficient of each monitoring point and all other monitoring points is taken as the vibration similarity coefficient of each monitoring point and other monitoring points.

[0086] The product of the average value of the vibration similarity coefficient of each detection point and the average value of the vibration stability coefficient is taken as the overall stability coefficient of the needle cylinder.

[0087] In other embodiments of the present application, the implementer can also use the weighted sum of the mean, median and mode to replace the average value, for example, to fuse the vibration similarity coefficients of all monitoring points with weights of 0.5, 0.2 and 0.3 to obtain the first stability factor; to fuse the vibration stability coefficients of all monitoring points with weights of 0.5, 0.2 and 0.3 to obtain the second stability factor; and to take the product of the first stability factor and the second stability factor as the overall stability coefficient of the needle cylinder.

[0088] It should be noted that both the Pearson correlation coefficient and the ReLU(x) function are prior art. In other embodiments of the present application, the DTW algorithm can be used to obtain the DTW similarity of the vibration data curves of different monitoring points to replace the Pearson correlation coefficient. The DTW algorithm is well known and will not be described here.

[0089] In the movement process of the circular knitting machine, when the stability of the needle cylinder is poor, it may be caused by the collision and friction between mechanical parts due to the unstable movement process of the circular knitting machine, or it may be caused by unstable voltage. Unstable voltage will cause motor speed fluctuation, and then make the rotation speed of the needle cylinder unstable. For example, instantaneous voltage drop or rise will cause the rotation speed of the needle cylinder to change in a short time, the responses of different point positions are different, and the vibration characteristics also change.

[0090] However, voltage mutation is usually caused by faults or abnormal operations of external power supply systems. Such sudden situations are rare, and there are devices such as power distribution cabinets that can stabilize voltage for timely control. In comparison, mechanical parts may collide and abnormally rub each other due to continuous friction, vibration and their own wear and tear during long-term operation. This wear and tear is a gradual accumulation process, and it is difficult to accurately predict when serious collision and friction problems will occur through simple monitoring means, making prevention difficult and affecting the production stability and quality of the circular knitting machine.

[0091] Therefore, according to the fluctuation characteristics of the voltage data in the preset historical neighborhood, the overall stability coefficient is combined to obtain a hardware abnormality factor, which avoids misjudging external voltage mutation as mechanical failure, so that the hardware abnormality factor can truly reflect the health status of the overall hardware of the circular knitting machine, and provide a reliable basis for dynamic switching of the control mode.

[0092] Preferably, in an embodiment of the present application, voltage fluctuations are generally manifested as sharp fluctuations near the extreme points, so observing the difference between adjacent amplitudes in the preset neighborhood of each voltage extreme point can effectively capture the local voltage instability characteristics; therefore, in the preset neighborhood of each voltage extreme point, the local instability coefficient of each voltage extreme point is obtained according to the difference between adjacent voltage amplitudes in time domain; and the unstable voltage points are screened based on the local instability coefficient;

[0093] Further considering that the smaller the number of unstable voltage points is, the smaller the overall stability coefficient is, which means that the instability of the circular knitting machine is more likely to be caused by hardware abnormalities, and the hardware abnormality factor is larger, so the number of unstable voltage points and the overall stability coefficient are fused to obtain the hardware abnormality factor, which can reflect the coupling of external power fluctuations and mechanical vibration states, thereby obtaining a more comprehensive and objective hardware abnormality factor.

[0094] As an example, the preset neighborhood is a time domain range of three seconds before and after the time corresponding to the voltage extreme point, and in the preset neighborhood of each voltage extreme point, the average value of the absolute value of the difference between adjacent voltage amplitudes in time domain is taken as the local instability coefficient of the corresponding voltage extreme point, the difference between adjacent voltage amplitudes in time domain is represented by the absolute value, and the fluctuation characteristics of the overall voltage data in the preset neighborhood are represented by the average value.

[0095] The larger the local instability coefficient is, the more unstable the voltage is, so the local instability coefficient is linearly normalized in the corresponding data dimension, a preset instability threshold 0.7 is set, and when the normalized local instability coefficient is greater than the preset instability threshold, the corresponding voltage extreme point is marked as an unstable voltage point.

[0096] The product of the number of unstable voltage points and the overall stability coefficient is added to the reciprocal of the sum value after the preset zero positive parameter 0.01, and then linearly normalized in the corresponding data dimension, as the hardware abnormality factor.

[0097] Wherein, the logical relationship is adjusted by taking the reciprocal, and the preset zero positive parameter is used to prevent the denominator from being zero; in other embodiments of the present application, the implementer can adjust the preset instability threshold.

[0098] Step S4: switching the control mode of the circular knitting machine based on the hardware abnormality factor.

[0099] The hardware abnormality factor provides a basis for judging the hardware abnormality of the circular knitting machine, and finally switches the control mode of the circular knitting machine based on the hardware abnormality factor.

[0100] Preferably, when the hardware abnormality factor is greater than a preset abnormality threshold, the fuzzy control is switched to control the circular knitting machine, and otherwise, the original control mode remains unchanged.

[0101] As an example, the preset abnormal threshold is 0.8, the fuzzy control is the prior art and is not the focus of the present scheme, and is briefly described herein, the spindle speed of the machine, the vibration amplitude of the needle cylinder at different points, the weight of the needle cylinder vibration data, the yarn tension, the voltage and the like are taken as the input data, and the execution parameters for adjusting the stability, such as the spindle speed adjustment amount, the yarn feeding speed adjustment amount and the like, are taken as the output data;

[0102] The execution mechanism of the equipment is controlled to change the unstable state, the machine is kept stable during operation, the vibration caused by factors such as mechanical part collision, friction and voltage mutation is reduced, the vibration stability and vibration similarity of each point of the needle cylinder are ensured to meet the normal operation requirements, and then the high-quality fabric is ensured to be produced by the circular knitting machine.

[0103] When the hardware abnormal factor is greater than the preset abnormal threshold, the excess production is carried out, that is, the feeding is stopped, the excess production is carried out after the excess production is completed, and the maintenance is carried out.

[0104] In another embodiment of the present application, the implementer can also issue a warning signal when the hardware abnormal factor is greater than the preset abnormal threshold, the relevant personnel take over the control of the circular knitting machine, and the relevant personnel carry out the emergency shutdown or the control of the excess production.

[0105] In other embodiments of the present application, the implementer can adjust the preset abnormal threshold.

[0106] One embodiment of the present application further provides a real-time control system for the motion stability of a circular knitting machine, which comprises a memory, a processor and a computer program, wherein the memory is used for storing the corresponding computer program, the processor is used for running the corresponding computer program, and the computer program can realize the real-time control method for the motion stability of the circular knitting machine when running in the processor.

[0107] In summary, in view of the technical problems that the hardware wear of the circular knitting machine is difficult to predict and maintain in time, and affects the production stability and production quality of the circular knitting machine, the present application provides a real-time control method and system for motion stability of a circular knitting machine. The present application divides the vibration period into normal period and abnormal period by extracting the fluctuation feature vectors of the vibration data in the vibration period and using the similarity between the fluctuation feature vectors. Further, according to the distribution of the normal period and the abnormal period, and combining the similarity of the fluctuation feature vectors between the abnormal periods, the vibration stability coefficient of each monitoring point at present is obtained. Further, according to the fluctuation similarity of the vibration data between the monitoring points, combining the distribution similarity of the normal period and all the vibration stability coefficients, the overall stability coefficient of the needle cylinder is obtained. Further, according to the fluctuation intensity feature of the voltage data in the preset historical neighborhood, combining the overall stability coefficient, the hardware abnormal factor is obtained. Finally, based on the hardware abnormal factor, the control mode of the circular knitting machine is switched. The present application analyzes the vibration data of each monitoring point, obtains the vibration period, the stability coefficient and the overall stability coefficient, and obtains the hardware abnormal factor by combining the voltage fluctuation feature, so that the hardware abnormality can be accurately and comprehensively identified, the control and maintenance can be switched in time, and the operation stability and production quality are improved.

[0108] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0109] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment mainly describes the differences from other embodiments.

Claims

1. A method for real-time control of motion stability on a circular knitting machine, characterized in that, The method includes: Obtain vibration and voltage data from each monitoring point of the circular knitting machine; within the current preset historical neighborhood, obtain the vibration period of each monitoring point based on the frequency domain distribution of the vibration data; Extract the wave feature vector of the vibration data within the vibration cycle; for each monitoring point, divide the vibration cycle into normal cycle and abnormal cycle according to the similarity between the wave feature vectors of different vibration cycles; based on the distribution of the normal cycle and the abnormal cycle, and combined with the similarity features of the wave feature vectors between the abnormal cycles, obtain the vibration stability coefficient of each monitoring point at the current time. Based on the similarity of vibration data fluctuations between monitoring points, combined with the similarity of normal cycle distribution and all vibration stability coefficients, the overall stability coefficient of the syringe is obtained; based on the violent fluctuation characteristics of voltage data within a preset historical neighborhood, combined with the overall stability coefficient, the hardware anomaly factor is obtained. The control mode of the circular knitting machine is switched based on the aforementioned hardware anomaly factor.

2. The method for real-time control of motion stability on a circular knitting machine according to claim 1, characterized in that, The method for obtaining the fluctuation feature vector includes: The amplitude corresponding to the maximum value of the vibration data in each cycle, and the time interval between the first trough and the first peak, constitute the wave characteristic vector.

3. The method for real-time control of motion stability on a circular knitting machine according to claim 2, characterized in that, The method for dividing the vibration period into normal and abnormal periods includes: At the same monitoring point, the average of the Euclidean distances between the wave characteristic vector of each vibration cycle and the wave characteristic vectors of all other vibration cycles is used as the vector difference coefficient of each vibration cycle; vibration cycles with vector difference coefficients less than a preset vector difference threshold are marked as normal cycles, and the remaining vibration cycles are marked as abnormal cycles.

4. The method for real-time control of motion stability on a circular knitting machine according to claim 1, characterized in that, The method for obtaining the vibration stability coefficient includes: At the same monitoring point, the fluctuation feature vectors of the abnormal cycles are clustered; within each cluster, the important influencing factors of each cluster are obtained based on the distribution of the abnormal cycles. Based on the number of normal cycles, the temporal proportion of the maximum continuous duration of the normal cycles, and all the important influencing factors, the vibration stability coefficient of the monitoring point is obtained.

5. The method for real-time control of motion stability on a circular knitting machine according to claim 4, characterized in that, The methods for obtaining the important impact factors include: Within each cluster, the key influencing factor of each cluster is obtained based on the minimum time interval between the time-adjacent abnormal cycles, the maximum time interval between the first and last two abnormal cycles in the time domain, and the number of abnormal cycles within the cluster.

6. The method for real-time control of motion stability on a circular knitting machine according to claim 1, characterized in that, The method for obtaining the overall stability coefficient includes: Obtain the Pearson correlation coefficient of the vibration data from any two monitoring points; The proportion of the number of normal cycles that overlap in the time domain of any two monitoring points is used as the similarity correction coefficient for the corresponding two monitoring points. By integrating the Pearson correlation coefficient and the similarity correction coefficient of each monitoring point with all other monitoring points, and combining all the vibration stability coefficients, the overall stability coefficient of the syringe is obtained.

7. The method for real-time control of motion stability on a circular knitting machine according to claim 1, characterized in that, The method for obtaining the hardware anomaly factor includes: Within a preset neighborhood of each voltage extreme point, a local instability coefficient is obtained for each voltage extreme point based on the difference in voltage amplitude between adjacent points in the time domain; unstable voltage points are then selected based on the local instability coefficient. By combining the number of unstable voltage points and the overall stability coefficient, the hardware anomaly factor is obtained.

8. The method for real-time control of motion stability on a circular knitting machine according to claim 1, characterized in that, The method for obtaining the vibration period includes: Perform a Fourier transform on the vibration data of each monitoring point in the current preset historical neighborhood, obtain the frequency with the highest amplitude in the spectrum as the dominant frequency, obtain the period length based on the dominant frequency, divide the preset historical neighborhood with the period length corresponding to each monitoring point, and obtain the vibration period of each monitoring point.

9. The method for real-time control of motion stability on a circular knitting machine according to claim 1, characterized in that, The method for switching the control mode of a circular knitting machine based on the hardware anomaly factor includes: When the hardware anomaly factor exceeds the preset anomaly threshold, the control of the circular knitting machine is switched to fuzzy control.

10. A real-time control system for motion stability of a circular knitting machine, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the real-time motion stability control method for a circular knitting machine as described in any one of claims 1 to 9.

Images