Multi-axis linkage winding quality monitoring device based on numerical control fiber winding

Abstract

The present application relates to the thickness quality monitoring technical field of winding process, specifically relates to multi-axis linkage winding quality monitoring equipment based on numerical control fiber winding, including memory and processor, the processor executes the computer program stored in the memory to realize: the winding thickness data of different positions in the axial direction of the material to be analyzed at each time is obtained;According to the difference between the winding thickness data of each position and adjacent position at each time, the local position segment is obtained;According to the difference of the winding thickness data corresponding to each local position segment and other local position segments at each time and the difference between the winding thickness data and the normal state, the winding abnormal value is obtained;Based on the winding abnormal value of each local position segment at each time, the fiber winding forming process is monitored, and the winding quality monitoring result is obtained. The present application improves the precision of the thickness abnormal detection result in the implementation monitoring process, so that the analysis result of thickness quality monitoring is more optimal.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of thickness quality monitoring of winding process, and in particular to a multi-axis linkage winding quality monitoring device based on numerical control fiber winding. BACKGROUND

[0002] Fiber winding forming is a high-efficiency manufacturing process commonly used for preparing composite material pole towers (such as conical electric poles, communication tower poles, etc.). The process usually adopts numerical control winding equipment to realize the precise superposition of fibers on the surface of a mold at a set angle (such as ± θ, 0°, 90°, etc.) by controlling the axial movement and rotation of the winding components. With the development of equipment performance and material systems, the manufacturing of composite pole towers gradually introduces the strategy of layered winding of structural layers and weather-resistant layers, cooperates with steam internal heating and cooling demolding processes, and significantly improves the strength, weather resistance, and manufacturing efficiency of the products. In actual production, to ensure winding quality, the numerical control program is usually relied on to control winding angles, speeds, tension, and other process parameters, and devices such as tension sensors are used to maintain winding stability.

[0003] In the layered winding manufacturing process of composite material pole towers, the structural layer and the weather-resistant layer are usually wound in sequence and bear the load and protection functions, respectively. Small fluctuations in the winding quality of the structural layer often amplify in the subsequent winding of the weather-resistant layer, thereby affecting the appearance consistency and service performance of the entire pole tower. The traditional numerical control winding equipment mainly relies on program settings and limited tension control means to monitor the quality of the entire material for abnormalities, lacks real-time sensing ability for actual thickness changes during the winding process, and is difficult to timely find local thickness deviation problems in the structural layer and to obtain the local position where the winding quality problem exists. SUMMARY

[0004] In order to solve the technical problem that the existing quality monitoring method relies on the setting means to monitor the entire winding material for abnormalities, resulting in low accuracy of thickness abnormality monitoring results, the purpose of the present application is to provide a multi-axis linkage winding quality monitoring device based on numerical control fiber winding, and the technical scheme adopted is as follows:

[0005] The multi-axis linkage winding quality monitoring device based on numerical control fiber winding comprises a storage and a processor, and the processor executes the computer program stored in the storage to realize the following steps:

[0006] In the fiber winding forming process, the winding thickness data of different positions in the axial direction of the material to be analyzed at each time is obtained;

[0007] segment each position of the material to be analyzed in the axial direction at each time according to the difference between the winding thickness data of each position and the winding thickness data of the adjacent positions in the neighborhood of the position distribution, to obtain each local position segment at each time;

[0008] obtain the winding abnormal value of each local position segment at each time according to the difference between the winding thickness data of each local position segment at each time and the winding thickness data of other local position segments, and the difference between the winding thickness data of each local position segment at each time and the normal state;

[0009] monitor the fiber winding forming process based on the winding abnormal value of each local position segment at each time, and obtain the winding quality monitoring result.

[0010] Preferably, the segmenting each position of the material to be analyzed in the axial direction at each time according to the difference between the winding thickness data of each position and the winding thickness data of the adjacent positions in the neighborhood of the position distribution, to obtain each local position segment at each time, specifically comprises:

[0011] obtain the mutation degree of each position at each time according to the difference between the winding thickness data of one side of each position in the axial direction and the winding thickness data of the other side;

[0012] segment all positions in the axial direction of the material to be analyzed by taking the position corresponding to the maximum point of the mutation degree of all positions at each time as a segment point, to obtain each local position segment at each time.

[0013] Preferably, the obtaining the mutation degree of each position at each time according to the difference between the winding thickness data of one side of each position in the axial direction and the winding thickness data of the other side, specifically comprises:

[0014] For any one time, taking any one position as a target position, the winding thickness data of the target position and the winding thickness data of the adjacent preset number of positions on the left side of the target position are obtained to form a first side data sequence of the target position;

[0015] the winding thickness data of the target position and the winding thickness data of the adjacent preset number of positions on the right side of the target position are obtained to form a second side data sequence of the target position;

[0016] determine the mutation degree of the target position at the any one time based on the difference between the mean value of all data in the first side data sequence and the mean value of all data in the second side data sequence.

[0017] Preferably, the difference between the winding thickness data corresponding to each local position segment and other local position segments at each time point, and the difference between the winding thickness data of each local position segment at each time point and the normal state, are used to obtain the winding abnormal value of each local position segment at each time point, specifically including:

[0018] The local position segments at each time point are clustered according to the difference distance between the winding thickness data corresponding to each local position segment and other local position segments at each time point, to obtain the feature cluster and the normal cluster at each time point.

[0019] The winding abnormal value of each local position segment in each feature cluster at each time point is obtained according to the difference between the winding thickness data contained in each feature cluster and the normal cluster.

[0020] Preferably, the local position segments at each time point are clustered according to the difference distance between the winding thickness data corresponding to each local position segment and other local position segments at each time point, to obtain the feature cluster and the normal cluster at each time point, specifically including:

[0021] For any one time point, the winding thickness data of all positions in each local position segment is used to form a thickness data sequence of each local position segment.

[0022] The DTW distance between the thickness data sequences of any two different local position segments is used as the metric distance between the two local position segments, and a clustering algorithm is used to cluster all local position segments to obtain a plurality of clustering clusters.

[0023] The difference between the average of the metric distance between any two local position segments in each clustering cluster and the average of the metric distance between any two local position segments in each other clustering cluster is determined as the data difference degree between each clustering cluster and each other clustering cluster.

[0024] The clustering cluster corresponding to the minimum value of the accumulation sum of the data difference degrees between each clustering cluster and all other cluster clusters is obtained as the normal cluster at the arbitrary time point, and all clustering clusters other than the normal cluster are used as the feature cluster at the arbitrary time point.

[0025] Preferably, the winding abnormal value of each local position segment in each feature clustering cluster at each time point is obtained according to the difference between the winding thickness data contained in each feature cluster and the normal cluster, specifically including:

[0026] For any one time point, the normalized result of the data difference degree between each feature cluster and the normal cluster is used as the winding abnormal value of each local position segment in each feature cluster.

[0027] Preferably, the winding abnormal value based on each local position segment at each time point is used to monitor the fiber winding forming process, specifically including:

[0028] When the winding abnormal value of each local position segment at each time point is greater than the preset abnormal threshold value, the local position segment contains a position with winding abnormal condition;

[0029] When the winding abnormal value of each local position segment at each time point is less than or equal to the preset abnormal threshold value, the local position segment contains a position without winding abnormal condition.

[0030] Preferably, after obtaining the winding quality monitoring result, it further includes:

[0031] Obtaining actual tension data of different positions in the axial direction of the material to be analyzed at each time point;

[0032] When the local position segment has winding abnormal condition, the actual tension data is corrected according to the difference between the winding thickness data and the normal state data of each position in the local position segment, combined with the winding abnormal value, to obtain the adjusted tension data of each position at each time point.

[0033] The next round of fiber winding operation is performed with the adjusted tension data, and the winding abnormal value of each local position segment at each time point in the next round of operation is obtained to monitor the fiber winding forming process.

[0034] Preferably, the actual tension data is corrected according to the difference between the winding thickness data and the normal state data of each position in the local position segment, combined with the winding abnormal value, to obtain the adjusted tension data of each position at each time point, specifically including:

[0035] When the local position segment has winding abnormal condition, the correction coefficient corresponding to the local position segment is obtained according to the difference between the winding thickness data contained in the feature cluster class and the normal cluster class of the local position segment at the corresponding time point, combined with the winding abnormal value of the feature cluster class where the local position segment is located; the product of the correction coefficient and the actual tension data of each position in the local position segment is taken as the corrected tension data.

[0036] Preferably, the correction coefficient corresponding to the local position segment is obtained according to the difference between the winding thickness data contained in the feature cluster class and the normal cluster class of the local position segment at the corresponding time point, combined with the winding abnormal value of the feature cluster class where the local position segment is located, specifically including:

[0037] When judging whether the average value of the winding thickness data of all positions in the feature cluster class where the local position segment is located is less than the average value of the winding thickness data of all positions in the normal cluster class.

[0038] If yes, the difference between the value 1 and the winding abnormal value of the feature cluster class in which the local position segment is located is taken as the correction coefficient corresponding to the local position segment;

[0039] If no, the sum of the value 1 and the winding abnormal value of the feature cluster class in which the local position segment is located is taken as the correction coefficient corresponding to the local position segment.

[0040] The embodiments of the present application have at least the following beneficial effects:

[0041] The present application first monitors the winding thickness data of the material in the axial direction in the winding forming process in real time. Then, by the difference between the winding thickness data of each position and the adjacent position, the data mutation of each position at each time is analyzed, and then all positions are segmented to obtain each local position segment at each time. Positions with similar data distribution can be divided into a segment, and then the material to be analyzed is divided into different local regions for feature detection. Further, the first aspect analyzes the data distribution difference between different local position segments at the same time, and the second aspect analyzes the data difference between each local position segment at each time and the normal state. The analysis results of the two aspects are combined to evaluate the winding abnormal value of the thickness abnormality size that may exist in each local position segment at each time. The present application can specifically exist winding thickness abnormality in each local position by real-time monitoring of the winding thickness data of each position, can obtain the local position of the material to be analyzed in the winding process, improves the accuracy of the thickness abnormality detection result in the implementation of the monitoring process, and makes the analysis result of the quality monitoring of the material to be analyzed more optimal. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art and the advantages thereof, 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 also obtain other drawings according to these drawings without any creative effort.

[0043] Figure 1 is a step flow chart of a multi-axis linkage winding quality monitoring method based on numerical control fiber winding provided by the present application;

[0044] Figure 2 is a step flow chart of a method for obtaining each local position segment at each time provided by the present application;

[0045] Figure 3 is a step flow chart of a method for obtaining the winding abnormal value of each local position segment provided by the present application;

[0046] Figure 4 is a step flow chart of the feature cluster and normal cluster acquisition method provided by the present application. DETAILED DESCRIPTION

[0047] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined inventive objectives, the specific implementation, structure, features and effects of the multi-axis linkage winding quality monitoring equipment based on numerical control fiber winding according to the present application are described in detail as follows in combination with the drawings and preferred embodiments.

[0048] Before introducing the specific schemes provided by the embodiments of the present application, some terms in the present application are explained to facilitate understanding by those skilled in the art, and do not limit the terms used in the present application.

[0049] Numerical control winding machine position data: through the pulse feedback of the servo motor or the stepper motor, the numerical control winding machine is controlled to move along the axial direction, and the winding machine position is recorded in real time. Spindle rotation speed: the rotation speed of the conical mold core in the winding process. Winding speed: determined by the discharge speed of the numerical control winding machine; winding angle: calculated from the spindle speed and the carriage feed speed, not actually detected, relying on the program accuracy; winding tension: the tension is set and adjusted through the tension motor, and the tension size is monitored in real time through the tension sensor. Winding surface profile data: 3D line laser displacement sensor, acquiring the relative distance of winding at each position in real time, which can reflect the thickness condition of winding at each position. In order to obtain real-time winding condition data, the line laser displacement sensor is installed on the guide rail along the axial direction of the conical mold during the winding process, which can detect the thickness change of the tower surface in real time during the winding process.

[0050] In the following description, different "one embodiment" or "another embodiment" refers to different embodiments, which are not necessarily the same. In addition, specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0051] 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 the present application belongs.

[0052] The specific scheme of the multi-axis linkage winding quality monitoring equipment based on numerical control fiber winding provided by the present application is described in detail below in combination with the drawings. The multi-axis linkage winding quality monitoring equipment based on numerical control fiber winding comprises a memory and a processor, and the processor executes the computer program stored in the memory to realize the steps of a multi-axis linkage winding quality monitoring method based on numerical control fiber winding.

[0053] More specifically, please refer to Figure 1Fig. 1 shows a step flow chart of a multi-axis linkage winding quality monitoring method based on numerical control fiber winding according to an embodiment of the present application, which comprises the following steps:

[0054] In step S100, the winding thickness data of different positions on the axial direction of the material to be analyzed at each time is obtained during the fiber winding forming process.

[0055] As a specific example, the material to be analyzed is a tower, and the winding forming process of its conical mold is taken as an example for a simple introduction. In the winding forming process of the tower, the material wound in the inner layer is a structural layer material, and after winding, the inner conical mold is heated to a gel state, and then a weather-resistant material is further wound on the outer layer, and after heating and solidification, the winding is completed.

[0056] During the winding process, the line and the tension of the winding composite material during the winding process are set by the numerical control machine tool. In the winding process, the temperature field of the mold core is uneven in the process of internal heating (steam heating) and external cooling, which causes problems such as "bulging" or "depression" after winding. Based on this problem, the relative distance data of the winding surface is obtained in real time to achieve the purpose of thickness monitoring during winding.

[0057] Specifically, a guide rail is arranged directly above the center axis of the mold in parallel, and laser displacement sensors are arranged at different positions of the guide rail. The sensor position setting implementer can select according to the specific implementation scene, and the distance interval between adjacent positions is equal, which can cover the tower as much as possible to realize real-time monitoring of winding abnormal conditions by analyzing the thickness distribution of winding at multiple different positions. During the winding process of the tower, the relative distance between the sensor position and the tower can be collected by using the sensor as the winding rotation process proceeds.

[0058] In order to obtain the thickness performance of the winding process, first, before the winding starts, the relative distance between each sensor corresponding position and the initial mold surface is obtained by using the sensor to establish the initial reference distance of the thickness condition of each position. Then, the winding process is monitored in real time, and at each time, the relative distance between the corresponding position and the tower in the winding process is obtained by using the sensor as the actual relative distance. Finally, the difference between the actual relative distance of each position at each time and the initial reference distance is the winding thickness data of each position at each time, which is used to monitor the thickness distribution of the winding material at different positions on the axial direction in the winding process.

[0059] In step S200, according to the difference between the winding thickness data of each position at each time and the winding thickness data of the adjacent positions in the neighborhood of the position distribution, all positions on the axial direction of the material to be analyzed at each time are segmented to obtain each local position segmentation at each time.

[0060] It should be understood that, in the winding process, the winding angle corresponding to different winding processes is not fixed. When the winding angle is small, the winding direction is relatively perpendicular to the axial direction of the mold. In a winding process in which the winding device winds from one end of the mold to the other end, the thickness of the mold surface changes relatively uniformly. Conversely, when the winding angle is large, the winding material does not cover the entire mold after one winding, and the thickness increases periodically with the size of the winding angle. At a certain moment, the change in the winding thickness data varies periodically with the position. In order to combine and analyze positions with similar thickness distributions, the method for obtaining each local position segment at each moment can be implemented by steps S201 and S202.

[0061] In some embodiments, as shown in FIG. 2A, the method for obtaining each local position segment at each moment can be implemented by steps S201 and S202. Figure 2

[0062] Step S201: According to the difference between the winding thickness data on one side and the winding thickness data on the other side of each position at each moment, the degree of mutation of each position at each moment is obtained.

[0063] It should be understood that the present embodiment is implemented for monitoring the thickness in the winding process of the tower, and therefore the method is implemented at each moment for quality evaluation. It should be noted that the possibility of abnormality is small in the initial stage of the winding process. In some embodiments, the implementer can also set the winding process to perform winding work for a fixed period of time before performing the abnormal monitoring operation. Based on this, the present embodiment takes any moment as an example, and monitors the local position that may have a thickness anomaly in real time based on the thickness distribution of different positions at the same moment. For the convenience of understanding and description, the current moment is taken as an arbitrary moment in the subsequent steps.

[0064] Specifically, taking the left end point of the tower as the starting point, the change in the winding thickness data is obtained in the order of each position from left to right. The position sequence number corresponding to each position is taken as the horizontal coordinate, and the winding thickness data of each position at the same moment is taken as the vertical coordinate. A thickness fitting curve is obtained by curve fitting. The winding change trend of different positions at the same moment can be reflected on the curve. When the winding thickness data of a position before and after the position is large, the position on the thickness fitting curve shows a mutation point. It should be understood that the difference between the winding thickness data of the left and right adjacent positions at the mutation point is maximized. Based on this feature, the difference between the winding thickness data of each position at the current moment and the winding thickness data of the adjacent two sides is analyzed to analyze the possibility of data mutation of each position at the current moment.

[0065] ​In the first step, for any time point, any position is taken as a target position, and the winding thickness data of the target position and the preset number of positions adjacent to the left of the target position are taken to form a first side data sequence of the target position.

[0066] In the second step, the winding thickness data of the target position and the preset number of positions adjacent to the right of the target position are taken to form a second side data sequence of the target position.

[0067] As a specific example, the preset number is 5, which can be determined by the implementer according to the length of the tower and the number of positions. For the current time point, the first side data sequence and the second side data sequence each contain 6 winding thickness data.

[0068] In the third step, the mutation degree of the target position at the time point is determined based on the difference between the mean of all data in the first side data sequence and the mean of all data in the second side data sequence.

[0069] As a specific example, for the current time point, the absolute value of the difference between the mean of all winding thickness data in the first side data sequence of the target position at the current time point and the mean of all winding thickness data in the second side data sequence is taken as the mutation degree of the target position at the current time point. The mutation degree reflects the difference between the data on both sides of the target position, and represents the mutation size of the data at the target position.

[0070] In step S202, the position corresponding to the maximum point of the mutation degree of all positions at each time point is taken as a segmentation point, and all positions on the axis of the material to be analyzed are segmented to obtain each local position segmentation at each time point.

[0071] For example, when the position numbers are x1 and x2, they are the maximum points of the mutation degree, so the positions corresponding to x1 and x2 are taken as segmentation points. Thus, from the starting position to the position number x1 belongs to the same local position segmentation, the position number x1+1 to x2 belongs to the same local position segmentation, and the position number x2 to the end position belongs to the same local position segmentation. It should be noted that the method for obtaining the maximum point is a known technology, and will not be described in detail here. The segmentation point is the position where the winding thickness data changes.

[0072] In step S300, the winding abnormal value of each local position segmentation at each time point is obtained according to the difference between the winding thickness data of each local position segmentation and other local position segmentations at each time point, and the difference between the winding thickness data of each local position segmentation at each time point and the normal state.

[0073] Since the thickness at different positions may be abnormal due to various factors during winding, it is necessary to detect the abnormality of each local position segment. The winding thickness of different position segments of the tower surface needs to be distinguished from normal process fluctuations and real abnormalities. The abnormality is not only reflected in the deviation of single-point thickness, but also in the cooperative deviation relationship between local area and overall state. Based on this consideration, first, the difference distance of the winding thickness data between each local position segment is divided into a normal state position set and a position set that may have an abnormal state in a clustered manner, and then the deviation between the local and the overall can be measured to quantify the winding abnormality of each local position segment.

[0074] As a specific example, as shown in Figure 3 , the method for obtaining the winding abnormal value of each local position segment can be implemented by step S301 and step S302.

[0075] Step S301, according to the difference distance of the winding thickness data corresponding to each local position segment and other local position segments at each time, clustering all local position segments at each time to obtain feature clusters and normal clusters at each time.

[0076] In order to identify the spatial abnormal pattern of winding thickness, the state association of each local position segment at the current time needs to be established. Since the winding angle changes dynamically, i.e. the thickness is uniformly distributed at small angles and periodically fluctuates at large angles, directly comparing the global thickness data will mask the local abnormality. Therefore, by comparing the data distribution difference of different local position segments, the division operation of similar cases is realized.

[0077] More specifically, as shown in Figure 4 , the method for obtaining the feature clusters and normal clusters can be implemented by steps S3011 to S3014.

[0078] Step S3011, for any time, the winding thickness data of all positions in each local position segment forms a thickness data sequence of each local position segment.

[0079] Step S3012, the DTW distance of the thickness data sequence between each two different local position segments is taken as the measurement distance between each two local position segments, and a clustering algorithm is used to cluster all local position segments to obtain a plurality of clustering clusters.

[0080] Specifically, K-means clustering algorithm can be used for processing, and the number of clustering clusters can be calculated according to the elbow method or the contour coefficient, which are both known technologies and will not be described in detail here. At the current time, all local position segments contained in each clustering cluster show a similar winding thickness data change trend.

[0081] Step S3013, the difference between the mean of the metric distance between each two local position segments in each cluster and the mean of the metric distance between each two local position segments in each other cluster is determined as the data difference degree between each cluster and each other cluster.

[0082] At the current time, any one cluster is taken as a target cluster, and other clusters are taken as reference clusters. A metric distance can be calculated between each two different local position segments. The absolute value of the difference between the mean of all metric distances in the target cluster and the mean of all metric distances in each reference cluster is calculated to obtain the data difference degree between the target cluster and each reference cluster.

[0083] The data difference degree reflects the difference in the thickness difference data between different clusters. When there is a large data difference between the target cluster and each reference cluster, the possibility of abnormality of the target cluster is greater.

[0084] Step S3014, the cluster corresponding to the minimum of the sum of the data difference degrees between each cluster and all other clusters is taken as the normal cluster at the arbitrary time, and all other clusters except the normal cluster are taken as the feature cluster at the arbitrary time.

[0085] When a "bulge" or "depression" occurs, a cluster with a large difference from the remaining clusters will appear in different clusters. The cluster with the minimum sum of differences from the remaining clusters is selected to represent the normal thickness range. The normal cluster at the current time represents the normal distribution range of the winding thickness data at the current time. The feature cluster represents the data distribution set that may have abnormal thickness.

[0086] Step S302, according to the difference between the winding thickness data contained in each feature cluster and the normal cluster, the winding abnormal value of each local position segment in each feature cluster at each time is obtained.

[0087] Specifically, for an arbitrary time, the normalized result of the data difference degree between each feature cluster and the normal cluster is taken as the winding abnormal value of each local position segment in each feature cluster. The normalization method can be the maximum-minimum method, which will not be described in detail here.

[0088] More specifically, in the same way as step S3013, the data difference degree between each feature cluster class and the normal cluster class at the current time can be obtained, representing the difference between each feature cluster class and the normal thickness distribution range. The greater the difference, the greater the difference between each feature cluster class and the normal state, and the greater the possibility of abnormality of each feature cluster class, and the greater the value of the corresponding winding abnormal value.

[0089] Step S400, based on the winding abnormal value of each local position segment at each time, monitoring the fiber winding forming process to obtain the winding quality monitoring result.

[0090] The greater the value of the winding abnormal value of the feature cluster class where each local position segment is located at the current time, the greater the possibility of abnormal thickness of the corresponding feature cluster class. The smaller the value of the winding abnormal value of the feature cluster class where each local position segment is located at the current time, the smaller the possibility of abnormal thickness of the corresponding feature cluster class.

[0091] Further, when the winding abnormal value of the feature cluster class where each local position segment is located at each time is greater than the preset abnormal threshold value, the local position segment contains a position with a winding abnormal condition, and at this time the winding quality is poor. At the current time, it means that during the winding process, the feature cluster class greater than the abnormal threshold value contains a local position segment with a more serious bulge or depression, and the difference in thickness during the winding process is too large, and the winding needs to be stopped for further screening.

[0092] When the winding abnormal value of the feature cluster class where each local position segment is located at each time is less than or equal to the preset abnormal threshold value, the local position segment contains a position without a winding abnormal condition, and at this time the winding quality is good. At the current time, it means that during the winding process, the thickness difference of the local position segment meeting the threshold requirement is small, and the possibility of thickness abnormality is small, so there is no winding abnormal condition. As a specific example, the value of the abnormal threshold value is 0.7, which can be set by the implementer according to the specific implementation scenario.

[0093] So far, the main purpose of the present embodiment is to detect thickness abnormalities globally for the material being wound. In other embodiments, the implementer can perform further quality monitoring operations on the material with local abnormalities according to the specific implementation scenario, for example, the relevant professional staff can also be reminded to further check the positions with abnormalities.

[0094] In summary, the embodiment can detect the existence of abnormal winding thickness at each local position in real time, and can obtain the local position of the material to be analyzed with abnormal thickness during the winding process, thereby improving the accuracy of the thickness abnormality detection result during the monitoring process, and making the analysis result of the quality monitoring of the material to be analyzed more optimal. The winding can be stopped for the part with abnormal thickness, so that the relevant staff can investigate the abnormal situation and take timely measures to avoid affecting the subsequent winding process.

[0095] In some embodiments, considering that the outer layer winding is performed when the structural layer does not enter a proper gel state, poor interlayer bonding may occur, especially when steam internal heating structure is used, the temperature distribution is complex, and only a preset path and a preset tension size can be completed according to the program. As a result, stress concentration or wrinkles, delamination, bulges or depressions may occur due to excessive or insufficient tension. After the local thickness abnormality detection in step S400 is completed, the tension size of the local position with thickness abnormality can be dynamically adjusted, for example, when the winding abnormality value of a local position segment is large, it indicates that the abnormality is more likely to occur. When the thickness is greater than the normal range, the tension is increased on the basis of the original value to suppress the bulge or a small protrusion, and when the thickness is less than the normal range, the tension of the depression is reduced to prevent the depression from being aggravated.

[0096] Based on this, the following steps are further included after step S400:

[0097] In step S500, the actual tension data of different positions in the axial direction of the material to be analyzed at each time is obtained. When there is winding abnormality in a local position segment, the actual tension data is corrected according to the difference between the winding thickness data and the normal state data of each position in the local position segment, and the adjustment tension data of each position at each time is obtained in combination with the winding abnormality value. The next round of fiber winding operation is performed with the adjustment tension data, the winding abnormality value of each local position segment at each time in the next round of operation is obtained, and the fiber winding forming process is monitored.

[0098] First, the tension data of each position at each time is collected by a tension micro sensor. In actual operation, the tension data may be a fixed value set in advance. The embodiment aims to obtain the tension data at the time when the thickness abnormality is detected, as the data basis for subsequent fine tuning of the tension, so as to attempt to correct the abnormal condition by adjusting the tension in the case of slight abnormality. For example, the fixed tension data at the corresponding time can be directly obtained by a numerical control winding machine, or a micro sensor can be arranged at a position of a winding device such as a yarn guide head or a winding head of the winding process to collect the tension data of different positions.

[0099] Then, the method for obtaining the adjusted tension data of each position at each time point is specifically as follows: when there is a winding abnormal condition in the local position segment, a correction coefficient corresponding to the local position segment is obtained according to the difference between the winding thickness data contained in the feature cluster class and the normal cluster class at the corresponding time point, and in combination with the winding abnormal value of the feature cluster class in which the local position segment is located; and the product between the correction coefficient and the actual tension data of each position in the local position segment is taken as the corrected tension data.

[0100] The method for obtaining the correction coefficient is specifically as follows: when judging whether the mean value of the winding thickness data of all positions in the feature cluster class in which the local position segment is located is less than the mean value of the winding thickness data of all positions in the normal cluster class; if yes, the difference between the value 1 and the winding abnormal value of the feature cluster class in which the local position segment is located is taken as the correction coefficient corresponding to the local position segment; and if no, the sum of the value 1 and the winding abnormal value of the feature cluster class in which the local position segment is located is taken as the correction coefficient corresponding to the local position segment.

[0101] For the convenience of description, the current time is taken as an example for illustration, and it is assumed that the local position segment corresponding to the winding abnormal condition at the current time is the ith local position segment, and the adjusted tension data of the ith local position segment at the current time can be represented by the formula:

[0102]

[0103] wherein N i ′ represents the adjusted tension data of the ith local position segment at the current time, represents the mean value of all winding thickness data in the feature cluster class in which the ith local position segment is located at the current time, represents the mean value of the desired winding thickness data in the normal cluster class at the current time, W i represents the winding abnormal value of the feature cluster class in which the ith local position segment is located at the current time, N i represents the tension data corresponding to the ith local position segment at the current time, is a correction coefficient.

[0104] For the ith local position segment at the current time, when , it is indicated that the overall distribution of the winding thickness of the local position segment with the abnormal condition is larger than the normal thickness range, and it is further indicated that the ith local position segment at the current time has a convex thickness abnormal condition, so the tension of the position needs to be relatively increased, and then the value of is 1, and the correction coefficient is represented as 1+W i .

[0105] When At this time, the overall distribution of the winding thickness of the local position segment indicating the current abnormality is smaller than the normal thickness range, and thus it is indicated that the i-th local position segment currently has a thickness abnormality of a depression, and thus the tension of the position needs to be relatively reduced, and then The value of W is -1, and the correction coefficient is represented as 1-W i .

[0106] It should be noted that in the embodiment, the same tension is used for winding operation in a length corresponding to a continuously distributed local position, that is, the tension data of all positions in the same local position segment do not change. According to the same method, after obtaining the adjustment tension data of each local position segment with an abnormal condition at the current time, the tension of the winding process can be controlled by a PID control method.

[0107] Further, considering that a large amplitude of tension regulation affects the normal winding process and may even cause winding abnormalities, the normal tension data value range can be set, and when the adjustment tension data is not within the normal value range, the adjustment tension data is not used for regulation operation. More specifically, the minimum tension data of the winding process is used as the lower limit of the normal value range, and the maximum tension data allowed by the winding process is used as the upper limit of the normal value range. When the adjustment tension data is greater than the upper limit of the normal value range, the tension is not regulated by the adjustment tension data, and the tension is regulated by the upper limit of the normal value range. When the adjustment tension data is less than the lower limit of the normal value range, the tension can be regulated by the lower limit of the normal value range.

[0108] Further, the drum or depression in the actual situation has a gradual accumulation process. After the adjustment of the tension size ends, if the drum or depression is caused by the inner structure layer gel, the drum or depression condition will improve to a certain extent. For factors such as guide wheel or tension arm wear, the guide yarn track is offset, and the local area appears overlapping or layering phenomenon. At this time, the abnormality will be further superimposed, causing the thickness abnormality of part of the area in the winding process to further increase.

[0109] After a round of winding operation is performed by using the adjusted adjustment tension data, it is determined whether the local thickness abnormality condition still exists on the surface of the wound material according to the method of steps S100 to S400. If it exists, it indicates that the thickness abnormality condition cannot be solved by fine-tuning the tension, and thus the winding needs to be stopped, and a warning is given to enable relevant professionals to investigate and repair the abnormality.

[0110] So far, the main purpose of step S500 is to determine whether the thickness abnormality can be solved by controlling the tension of the position where the thickness abnormality exists, and this method can timely find the local position where the thickness abnormality exists in the early stage. If there is still a large abnormality after running for a period of time after dynamic parameter adjustment, relevant professionals need to further troubleshoot each local position.

[0111] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A multi-axis linkage winding quality monitoring device based on numerical control fiber winding, characterized in that, The device comprises a memory and a processor, the processor executes a computer program stored in the memory to realize the following steps: In the fiber winding forming process, the winding thickness data of different positions on the axial direction of the material to be analyzed at each time is obtained; According to the difference between the winding thickness data of each position at each time and the winding thickness data of adjacent positions in the neighborhood of the position distribution, all positions on the axial direction of the material to be analyzed at each time are segmented to obtain each local position segment at each time, including: For any one time, any one position is recorded as a target position, the winding thickness data of the target position and the preset number of adjacent positions on the left side of the target position are obtained to form a first side data sequence of the target position; The winding thickness data of the target position and the preset number of adjacent positions on the right side of the target position are obtained to form a second side data sequence of the target position; Based on the difference between the mean value of all data in the first side data sequence and the mean value of all data in the second side data sequence, the mutation degree of the target position at the arbitrary time is determined; The position corresponding to the maximum point of the mutation degree of all positions at each time is taken as a segmentation point, and all positions on the axial direction of the material to be analyzed are segmented to obtain each local position segment at each time; According to the difference between the winding thickness data of each local position segment and other local position segments at each time, and the difference between the winding thickness data of each local position segment at each time and the normal state, the winding abnormal value of each local position segment at each time is obtained; Based on the winding abnormal value of each local position segment at each time, the fiber winding forming process is monitored to obtain a winding quality monitoring result.

2. The numerically controlled fiber-winding based multi-axis linkage winding quality monitoring apparatus according to claim 1, characterized in that, The winding abnormal value of each local position segment at each time is obtained according to the difference between the winding thickness data of each local position segment and other local position segments at each time, and the difference between the winding thickness data of each local position segment at each time and the normal state, specifically including: According to the difference distance between the winding thickness data corresponding to each local position segment and other local position segments at each time, all local position segments at each time are clustered to obtain feature cluster classes and normal cluster classes at each time; According to the difference between the winding thickness data contained in each feature cluster class and the normal cluster class, the winding abnormal value of each local position segment in each feature cluster class at each time is obtained.

3. The numerically controlled fiber-winding based multi-axis linkage winding quality monitoring apparatus according to claim 2, characterized in that, The winding abnormal value of each local position segment at each time is obtained according to the difference distance between the winding thickness data corresponding to each local position segment and other local position segments at each time, and the difference between the winding thickness data of each local position segment at each time and the normal state, specifically including: For any one time, the winding thickness data of all positions in each local position segment forms a thickness data sequence of each local position segment; The DTW distance between the thickness data sequences of each two different local position segments is taken as the measurement distance between each two local position segments, and all local position segments are clustered by using a clustering algorithm to obtain a plurality of clustering clusters; differences between the mean of the metric distances between each two local position segments within each cluster and the mean of the metric distances between each two local position segments within each other cluster are determined as the data difference degree between each cluster and each other cluster; a cluster corresponding to the minimum of the accumulation sum of the data difference degrees between each cluster and all other cluster is obtained as the normal cluster at the arbitrary moment, and all other clusters except the normal cluster are taken as the feature clusters at the arbitrary moment.

4. The numerically controlled fiber-winding based multi-axis linkage winding quality monitoring apparatus according to claim 3, characterized in that, the winding abnormal value of each local position segment within each feature cluster at each moment is obtained according to the difference of the winding thickness data contained in each feature cluster and the normal cluster, and specifically includes: for an arbitrary moment, the normalization result of the data difference degree between each feature cluster and the normal cluster is taken as the winding abnormal value of each local position segment within each feature cluster.

5. The numerically controlled fiber-winding based multi-axis linkage winding quality monitoring apparatus according to claim 2, wherein, the fiber winding forming process is monitored based on the winding abnormal value of each local position segment at each moment, and specifically includes: when the winding abnormal value of each local position segment at each moment is greater than a preset abnormal threshold, the position contained in the local position segment has a winding abnormal condition; when the winding abnormal value of each local position segment at each moment is less than or equal to the preset abnormal threshold, the position contained in the local position segment does not have a winding abnormal condition.

6. The numerically controlled fiber-winding based multi-axis linkage winding quality monitoring apparatus according to claim 2, wherein, after obtaining the winding quality monitoring result, it further includes: actual tension data of different positions in the axial direction of the material to be analyzed at each moment is obtained; when the local position segment has a winding abnormal condition, the actual tension data is corrected according to the difference between the winding thickness data and the normal state data of each position within the local position segment, and the adjusted tension data of each position at each moment is obtained; the next round of fiber winding operation is performed with the adjusted tension data, the winding abnormal value of each local position segment at each moment in the next round of operation is obtained, and the fiber winding forming process is monitored.

7. The numerically controlled fiber-winding based multi-axis linkage winding quality monitoring apparatus according to claim 6, characterized in that, the actual tension data is corrected according to the difference between the winding thickness data and the normal state data of each position within the local position segment, and the adjusted tension data of each position at each moment is obtained, and specifically includes: when the local position segment has a winding abnormal condition, a correction coefficient corresponding to the local position segment is obtained according to the difference between the winding thickness data contained in the feature cluster and the normal cluster to which the local position segment belongs at the corresponding moment, and the winding abnormal value of the feature cluster to which the local position segment belongs; the product between the correction coefficient and the actual tension data of each position in the local position segment is taken as the corrected tension data.

8. The numerically controlled fiber-winding based multi-axis linkage winding quality monitoring apparatus according to claim 7, characterized in that, the correction coefficient corresponding to the local position segment is obtained according to the difference between the winding thickness data contained in the feature cluster and the normal cluster to which the local position segment belongs at the corresponding moment, and the winding abnormal value of the feature cluster to which the local position segment belongs, and specifically includes: whether the average of the winding thickness data of all positions in the feature cluster class where the local position segment is located is less than the average of the winding thickness data of all positions in the normal cluster class; if yes, then the difference between the value 1 and the winding abnormal value of the feature cluster class where the local position segment is located is taken as the correction coefficient corresponding to the local position segment; if no, then the sum of the value 1 and the winding abnormal value of the feature cluster class where the local position segment is located is taken as the correction coefficient corresponding to the local position segment.

Images