Fault self-diagnosis method and system for full-automatic hollow blow molding equipment

Abstract

The invention relates to the technical field of blow molding, in particular to a fault self-diagnosis method and system for full-automatic hollow blow molding equipment. The method comprises the following steps: firstly, acquiring all wall thickness non-uniform areas in a light-transmitting image of each blow molding product during wall thickness detection, and further performing inter-frame matching on the wall thickness non-uniform areas in different light-transmitting images to obtain batch wall thickness non-uniform areas; then acquiring a process fluctuation coefficient according to the fluctuation change of the monitoring curve of each process parameter in the blow molding process of the blow molding product; and finally, according to the batch number and the process fluctuation coefficient of the blow molding products corresponding to the batch areas with the uneven wall thicknesses, in combination with the gray level distribution and area change characteristics in the areas with the uneven wall thicknesses, carrying out fault early warning on the blow molding equipment. According to the invention, by analyzing the batch condition and accumulation characteristics of the areas with uneven inner wall thickness of the blow molding product and combining the fluctuation change of the parameters in the technological process, the possibility that the blow molding product is caused by the equipment fault is comprehensively evaluated, so that fault self-diagnosis and early warning are accurately carried out on the blow molding equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of blow molding, in particular to a fault self-diagnosis method and system of a full-automatic hollow blow molding equipment. BACKGROUND

[0002] The full-automatic hollow blow molding equipment blows hollow plastic products by heating plastic raw materials and using compressed air. The uniformity of the wall thickness of the blow molding products has a great influence on the product quality such as strength and sealing performance, and abnormal failure of the blow molding equipment is one of the main reasons for the non-uniform wall thickness of the blow molding products. Therefore, it is crucial to evaluate the failure of the full-automatic hollow blow molding equipment during the blow molding process.

[0003] At present, the wall thickness uniformity of the blow molding products is mainly detected to evaluate the failure of the full-automatic hollow blow molding equipment, thereby triggering the related failure shutdown early warning mechanism. The blow molding pressure is also one of the main factors for controlling the uniform molding of the blow molding products. The accidental fluctuation of the blow molding pressure in the blow molding process may cause the strain rate of the parison to increase sharply, thereby causing the local wall thickness of the blow molding products to be non-uniform. Such normal process fluctuation interference may cause misdiagnosis of the equipment failure, thereby possibly causing misshutdown and affecting the production efficiency. SUMMARY

[0004] In order to solve the technical problem of poor fault diagnosis effect of the full-automatic hollow blow molding equipment, the purpose of the present application is to provide a fault self-diagnosis method and system of a full-automatic hollow blow molding equipment, and the technical solution adopted is as follows: A fault self-diagnosis method of a full-automatic hollow blow molding equipment, the method comprising: Until the current time, all wall thickness non-uniform areas in the light transmission image of each blow molding product in the same batch during wall thickness detection are obtained. The monitoring curves of each process parameter of each blow molding product during the blow molding process are obtained, and the process parameters at least include the blow molding pressure and the locking current. According to the position, area and gray scale distribution change of the wall thickness non-uniform area in the light transmission image, the wall thickness non-uniform areas in different light transmission images of the blow molding products are matched between frames, and each frame matching result is taken as a batch of wall thickness non-uniform areas. For each blow molding product, the process fluctuation coefficient is obtained according to the fluctuation change of the monitoring curve of each process parameter during the blow molding process. At the current time, according to the batch quantity of the blow molding products corresponding to each batch of wall thickness non-uniform areas and the process fluctuation coefficient, the gray scale distribution and the area change characteristics in the wall thickness non-uniform area are combined to perform failure warning on the blow molding equipment.

[0005] Further, the method for obtaining the wall thickness non-uniform area comprises: The continuous light transmission frame image and the infrared frame image of the blow molding product in the wall thickness detection conveying process are acquired, the continuous light transmission frame image includes a light transmission image; the suspected wall thickness uneven area and uneven characteristic parameters thereof are determined according to the local gray scale distribution in each light transmission frame image, and the temperature information of each suspected wall thickness uneven area is acquired by combining the pixel value in the infrared frame image with the same frame number. For each blow molding product, the suspected wall thickness uneven area is matched between frames according to the position, area and temperature information of the suspected wall thickness uneven area in the light transmission frame image, and the wall thickness uneven area in the corresponding light transmission frame image is determined based on the matching result.

[0006] Further, the method for acquiring the suspected wall thickness uneven area and uneven characteristic parameters thereof comprises: In each light transmission frame image, all target points are screened out according to the gradient amplitude of the pixel points, the target points are subjected to region connectedness detection, and the uneven characteristic parameters of each connected domain are acquired according to the gray scale range and the distribution characteristics of the gradient amplitude of the pixel points in each connected domain, and all suspected wall thickness uneven areas are screened out from all connected domains according to the uneven characteristic parameters.

[0007] Further, the method for acquiring the batch wall thickness uneven area comprises: Any light transmission image is taken as a target image, and any wall thickness uneven area in the target image is taken as a target area, and each wall thickness uneven area in each non-target image is taken as a reference area. The first matching parameter is acquired by fusing the center position difference and the area difference between the target area and each reference area, and the negative correlation mapping result of the difference between the uneven characteristic parameters of the corresponding areas is taken as the second matching parameter; the matching coefficients between each reference area and the target area are acquired by fusing the first matching parameter and the second matching parameter. In each non-target image, the reference area with the maximum matching coefficient and greater than a preset matching threshold between the target area is screened out as the matching area of the target area; the target area and all matching areas thereof are taken as a batch wall thickness uneven area.

[0008] Further, the method for acquiring the process fluctuation coefficient comprises: For each blow molding product, the monitoring curve of each process parameter is segmented by using the extreme point, and all abnormal growth sub-sections are screened out from all subsections according to the growth amplitude of each subsection; the process fluctuation coefficient is acquired according to the growth amplitude and the duration of all abnormal growth sub-sections in each monitoring curve.

[0009] Further, the method for acquiring the process fluctuation coefficient according to the growth amplitude and the duration of all abnormal growth sub-sections in each monitoring curve comprises: In each monitoring curve, the mean value of the growth amplitude of the abnormal growth sub-section is multiplied by the duration ratio of all abnormal growth sub-sections in the monitoring curve to obtain a fluctuation parameter; and the normalized result of the cumulative sum of the fluctuation parameters of all monitoring curves is taken as the process fluctuation coefficient.

[0010] Further, the method for fault early warning of the blow molding equipment comprises: The batch quantity of the blow molding product corresponding to each type of batch wall thickness uneven area is negatively correlated to obtain a process disturbance weight; and the discrete eigenvalue of the process fluctuation coefficient of the blow molding product is weighted and normalized by using the process disturbance weight to obtain a process disturbance coefficient of each type of batch wall thickness uneven area. According to the area of each type of batch wall thickness uneven area and the batch accumulation of the uneven characteristic parameters, a fault disturbance coefficient of each type of batch wall thickness uneven area is obtained, and a suspected fault affected area is determined from all types of batch wall thickness uneven areas in combination with the process disturbance coefficient. According to the distribution characteristics of the fault disturbance coefficient of the suspected fault affected area in each blow molding product, in combination with the total number and total area of the fault affected area, a production abnormality coefficient of the blow molding product is obtained; and when the production abnormality coefficient of a preset number of blow molding products is greater than a preset early warning threshold value, a fault of the blow molding equipment is early warned.

[0011] Further, the method for obtaining the fault disturbance coefficient comprises: For each type of batch wall thickness uneven area, the area and the uneven characteristic parameters are respectively constructed into corresponding change sequences based on the detection sequence of the blow molding product, and the area growth parameter and the characteristic growth parameter are respectively obtained according to the number and accumulation characteristics of the non-negative values in the first-order difference sequence of the change sequence; and the fault disturbance coefficient is obtained by fusing the area growth parameter and the characteristic growth parameter.

[0012] Further, the method for obtaining the suspected fault affected area comprises: The batch wall thickness uneven area with the fault disturbance coefficient less than or equal to a preset fault threshold value and the process disturbance coefficient greater than a preset process threshold value is taken as a suspected process affected area; and the remaining batch wall thickness uneven areas are taken as suspected fault affected areas.

[0013] The present application also provides a full-automatic hollow blow molding equipment fault self-diagnosis system, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the steps of the full-automatic hollow blow molding equipment fault self-diagnosis method when executing the computer program.

[0014] The present application has the following beneficial effects: The application firstly acquires all wall thickness uneven areas in the light transmission image of each blow molding product in the same batch at the wall thickness detection time, prepares for subsequent batch feature evaluation, and acquires the monitoring curve of each process parameter in the blow molding process of each blow molding product to evaluate the fluctuation change and acquire the process fluctuation parameter; further based on the batch feature of wall thickness unevenness caused by equipment failure, the wall thickness uneven areas in different light transmission images of the blow molding product are matched between frames according to the position, area and gray scale distribution change of the wall thickness uneven areas in the light transmission image, and all batch wall thickness uneven areas are acquired; finally, at the current time, the process interference possibility is analyzed according to the batch quantity of the blow molding product corresponding to each batch wall thickness uneven area and the process fluctuation coefficient, and the batch accumulation feature is analyzed in combination with the gray scale distribution and area change feature of the wall thickness uneven area, so as to analyze the possibility of equipment failure, and then the blow molding equipment is fault warned. The application analyzes the batch situation and batch accumulation feature of the wall thickness uneven area in the blow molding product, and combines the fluctuation change of the process parameter in the blow molding process, so as to comprehensively evaluate the possibility of equipment failure, thereby accurately performing fault self-diagnosis and warning on the blow molding equipment. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, below briefly introduces the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only show some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative effort based on these drawings.

[0016] Figure 1 A flow chart of a fault self-diagnosis method of a full-automatic hollow blow molding equipment provided by an embodiment of the present application; Figure 2 A flow chart of a batch wall thickness uneven area acquisition method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined invention purpose, below, combined with the drawings and preferred embodiments, the specific implementation, structure, features and effects of the fault self-diagnosis method and system of a full-automatic hollow blow molding equipment according to the present application are described in detail as follows. 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.

[0018] 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.

[0019] The application provides a fault self-diagnosis method and system of a full-automatic hollow blow molding equipment.

[0020] Please refer to Figure 1 which shows a flowchart of a fault self-diagnosis method of a full-automatic hollow blow molding equipment according to an embodiment of the application, and specifically comprises the following steps. Step S1, at the current time, all uneven wall thickness areas in a light transmission image of each blow molding product in the same batch at the time of wall thickness detection are acquired; and a monitoring curve of each process parameter of each blow molding product in the blow molding process is acquired, the process parameters at least including blow molding pressure and mold locking current.

[0021] In an embodiment of the application, before the current time, first, in the product quality inspection process of the same batch, a light transmission image of each blow molding product in all inspected products collected at the time of wall thickness detection is acquired, so as to analyze the uneven wall thickness subsequently. Wherein, the acquisition of the light transmission image of the blow molding product is an existing technology well known by those skilled in the art, and the acquisition process is briefly described as follows: in the wall thickness detection process section of the conveying belt, a backlight light source is installed on one side, and a wall thickness detection device, i.e. an industrial camera and an infrared imaging device, is installed on the other side, so that the detection device and the backlight light source are coaxial, when the blow molding product reaches the specified station, the light source, the product and the device are coaxial, and the detection device will automatically collect continuous frame light transmission frame images and continuous frame infrared frame images.

[0022] As an example, the frame rate of the industrial camera and the infrared imaging device is set to 30 fps, and the continuous frame images in one second are collected, and the implementer can also adjust the acquisition process; and then the first frame light transmission frame image is taken as the light transmission image.

[0023] It should be noted that the blow molding product is detected one by one, that is, the light transmission image only contains one blow molding product.

[0024] After the light transmission image of each blow molding product is acquired, the uneven wall thickness area in the blow molding product can be further analyzed and determined, so as to analyze whether it is caused by equipment failure for subsequent analysis, thereby giving a warning.

[0025] Specifically, first, the corresponding area of the blow molding product in each light transmission image is determined by semantic recognition, and a threshold segmentation algorithm can also be used, so as to analyze and determine all uneven wall thickness areas in the blow molding product; and they are all known technologies, and will not be described here.

[0026] Considering that the blow molding product is usually detected for wall thickness after blow molding demolding, it still remains a certain temperature during conveying, the temperature distribution of the wall thickness uneven area is obviously different from that of the wall thickness uniform area, and the position of the wall thickness uneven area in the continuous frame image only changes slightly during the conveying of the blow molding product; and considering that the product wall thickness is different, the light transmittance is also different, the wall thickness uneven area will be characterized by uneven texture distribution in the light transmission image; Based on this, in one preferred embodiment of the present application, the method for obtaining the wall thickness uneven area comprises: obtaining continuous light transmission frame images and infrared frame images of the blow molding product during the wall thickness detection conveying process, the continuous light transmission frame images comprising light transmission images; determining a suspected wall thickness uneven area and uneven characteristic parameters thereof according to the local gray scale distribution in each light transmission frame image, and obtaining temperature information of each suspected wall thickness uneven area according to the pixel value in the infrared frame image of the same frame number; For each blow molding product, the suspected wall thickness uneven area is matched between frames according to the position, area and temperature information of the suspected wall thickness uneven area in the light transmission frame image where it is located, and the wall thickness uneven area in the corresponding light transmission frame image is determined based on the matching result.

[0027] In one preferred embodiment of the present application, considering that the wall thickness unevenness is mainly reflected in the disorder of gray scale texture change, and the gradient can reflect the gray scale change rate, the method for obtaining the suspected wall thickness uneven area and uneven characteristic parameters thereof comprises: In each light transmission frame image, all target points are selected according to the gradient amplitude of the pixel points, the region connectedness detection is performed on the target points, the uneven characteristic parameters of each connected domain are obtained according to the distribution characteristics of the gray scale range and the gradient amplitude of the pixel points in each connected domain, and all suspected wall thickness uneven areas are selected from all connected domains according to the uneven characteristic parameters.

[0028] As an example, in each light transmission frame image, the gradient of each pixel point is obtained based on the Sobel operator, then the pixel points with gradient amplitude greater than the average level are selected as target points, the region connectedness detection is performed on the target points, all connected domains are obtained, and the implementer can also perform certain inflation processing on each connected domain to improve the region integrity; then the product of the gray scale range and the average value of the gradient amplitude of all pixel points in each connected domain is obtained, the product is linearly normalized to obtain the uneven characteristic parameters of each connected domain; and all connected domains with uneven characteristic parameters greater than a preset threshold value such as 0.5 are selected from all connected domains as suspected wall thickness uneven areas. It should be noted that the linear normalization is performed in the dimension of all connected domains in each light transmission frame image, which is already a prior art together with the gradient acquisition, connected domain detection and inflation processing, and will not be described herein; the implementer can also adjust the preset threshold value or adjust the selection condition of the target points; After screening each suspected wall thickness area, the temperature information of each suspected wall thickness area can be further determined by synchronously collecting infrared frame graphs under the same frame serial number; specifically, the infrared pixel values of all pixel points in the same position area of the suspected wall thickness area in the infrared frame graph can be obtained by the mask method, the range of the infrared pixel values is taken as the temperature information, i.e. the temperature range of the suspected wall thickness uneven area, and the temperature uneven characteristics of the suspected wall thickness area are reflected; Further, in the continuous frame graphs of each blow molding product, the image coordinate system is constructed with the image center as the origin, the position coordinates of the region center of each suspected wall thickness uneven area in the corresponding frame graph are determined, the total number of pixel points in each suspected wall thickness uneven area is taken as the area, and the feature triple is constructed in combination with the temperature range of each suspected wall thickness uneven area; In the continuous light transmission frame graphs, the first frame is taken as the reference, any suspected wall thickness uneven area in the first frame is taken as the matching target, the feature difference (i.e. the Euclidean norm of the corresponding position coordinates, the area and the temperature range, and the Euclidean norm is mapped into the value range of tanh) between the matching target and the corresponding feature triple of each suspected wall thickness uneven area in the remaining frame graphs is obtained, the suspected wall thickness uneven area with the smallest feature difference and smaller than a preset threshold such as 0.2 in each frame is screened out, and the suspected wall thickness uneven area is taken as the matching area of the matching target in the frame. The above operation means is a prior art known to those skilled in the art, and will not be described in detail.

[0029] It should be noted that all parameters in the embodiments of the present application are uniformly subjected to standardization processing before operation, and the dimension influence is removed.

[0030] It should be noted that the matching target may not exist in the remaining frame graphs, when the number of matching areas is greater than a preset proportion such as 70% of the total number of continuous frame graphs, it is determined that the matching target is a wall thickness uneven area, and the implementer can also adjust the preset proportion.

[0031] At this point, the suspected wall thickness uneven area is matched between frames, and all wall thickness uneven areas in the light transmission image of each blow molding product are determined.

[0032] In another embodiment of the present application, all wall thickness uneven areas can also be directly determined by using a pre-trained labeling model, which is a prior art and will not be described in detail.

[0033] Considering that in the blow molding production process, the fluctuation of the blow molding process, such as the random fluctuation factors of the blow molding pressure due to unstable gas supply, will also cause quality problems of the blow molding product and the problem of uneven wall thickness, in order to further screen out the process influence and accurately evaluate the equipment failure, the monitoring curves of each process parameter in the blow molding process of each blow molding product are further obtained in the embodiments of the present application. Wherein, considering that the mold remains firmly closed during the blow molding process, when the blow molding pressure fluctuates, the gas pressure inside the mold and the force of the plastic expansion are also fluctuating, thereby affecting the closing force required by the mold locking system, and further causing the mold locking current to change; based on this, the process parameters at least include the blow molding pressure and the mold locking current, and the implementer can also increase other process fluctuation related process parameters.

[0034] Then the blow molding process of each blow molded product is traced back to obtain the monitoring curve of each process parameter in the corresponding blow molding process, that is, the blow molding pressure monitoring curve and the mold locking current monitoring curve; it should be noted that tracing and obtaining the monitoring curve are prior art means and will not be described in detail.

[0035] Step S2, according to the position, area and gray scale distribution change of the uneven wall thickness area in the light transmission image, the uneven wall thickness areas in different light transmission images of the blow molded product are matched between frames, and each frame matching result is taken as a batch of uneven wall thickness areas.

[0036] Considering that accidental process fluctuations and equipment abnormal failures in the same batch of blow molded products will cause batch problems of blow molding quality, the uneven wall thickness problem will batch appear in different blow molded products; the batch problem caused by accidental process fluctuations is random and less, and the uneven wall thickness batch caused by equipment abnormal failure is large and may have cumulative characteristics; and considering that the wall thickness detection process of the same batch of blow molded products is consistent, when the position, area and gray scale texture of the uneven wall thickness area are relatively similar or consistent, it means that it is more likely to be a batch problem; based on this, the uneven wall thickness areas in different light transmission images of the blow molded product can be matched between frames based on the similar logic of the frame matching of the suspected uneven wall thickness area in step S1 to determine the batch uneven wall thickness area, and prepare for subsequent evaluation of equipment abnormalities.

[0037] In a preferred embodiment of the present application, the method for obtaining the batch uneven wall thickness area comprises: Please refer to Figure 2 which shows a method flow chart for obtaining a batch uneven wall thickness area provided by an embodiment of the present application, and specifically comprises: Step S201, taking any light transmission image as a target image, and taking any uneven wall thickness area in the target image as a target area, and taking each uneven wall thickness area in each non-target image as a reference area.

[0038] First, determine the target area in the target image, take the target area as an example for matching expression, and then change the target image and the target area, so that the uneven wall thickness area can be matched between frames.

[0039] Step S202, between the target region and each reference region, fuse the center position difference between the regions and the area difference to obtain a first matching parameter, and map the negative correlation of the difference between the uneven feature parameters of the corresponding regions to a second matching parameter; fuse the first matching parameter and the second matching parameter to obtain a matching coefficient between each reference region and the target region.

[0040] As an example, between the target region and each reference region, the sum of the Euclidean distance between the position coordinates of the region centers and the absolute value of the area difference is mapped into an exponential function exp(-x) with a natural constant e as the base number to adjust the logic, and the mapping value is taken as the first matching parameter; the absolute value of the difference between the uneven feature parameters of the regions is also mapped into exp(-x) to adjust the logic, and the mapping value is taken as the second matching parameter; finally, the first matching parameter and the second matching parameter are multiplied to obtain a matching parameter, so that the smaller the area difference, position difference and uneven feature parameter difference of the two regions, the larger the matching parameter.

[0041] In other examples, the implementer can also adjust the negative correlation mapping method.

[0042] Step S203, in each non-target image, screen the reference region with the maximum matching coefficient and greater than a preset matching threshold from the target region as the matching region of the target region; and take the target region and all the matching regions thereof as a batch of wall thickness uneven regions.

[0043] As an example, in each non-target image, the preset matching threshold is set to 0.7 to obtain the matching region of the target region; and between all the blow molding product transmission images, the target region and all the matching regions thereof are taken as a batch of wall thickness uneven regions.

[0044] It should be noted that there is a possibility that the wall thickness uneven region in part of the transmission image has no matching result, and this wall thickness uneven region may be caused by accidental process fluctuation and does not have batch characteristics.

[0045] Step S3, for each blow molding product, according to the fluctuation change of the monitoring curve of each process parameter in the blow molding process, a process fluctuation coefficient is obtained; at the current time, according to the batch quantity of the blow molding product corresponding to each batch of wall thickness uneven regions and the process fluctuation coefficient, combined with the gray distribution and area change characteristics in the wall thickness uneven region, a fault warning is performed on the blow molding equipment.

[0046] Considering that the fluctuation of the process parameters will affect the diagnosis accuracy of the equipment fault, based on this, the embodiment of the application first evaluates the process fluctuation coefficient in the blow molding process of each blow molding product, and the process fluctuation coefficient is used to evaluate the possibility that the batch wall thickness uneven region is caused by process fluctuation interference, and to prepare for accurate evaluation of equipment abnormalities.

[0047] Preferably, in an embodiment of the present application, the sudden increase amplitude of the process parameter reflects the fluctuation degree, and the more frequent and longer the sudden increase, the more frequent and intense the fluctuation, which reflects the greater possibility of being the cause of uneven wall thickness; based on this, the method for obtaining the process fluctuation coefficient comprises: For each blow molding product, the monitoring curve of each process parameter is segmented by using the extreme points, and all abnormal growth segments are selected from all segments according to the growth amplitude of each segment; the process fluctuation coefficient is obtained according to the growth amplitude and the duration of all abnormal growth segments in each monitoring curve.

[0048] As an example, taking any blow molding product as an example, the extreme points in each monitoring curve are obtained, the extreme points are taken as segmentation points, the curve segment between adjacent extreme points is taken as a segment, the difference between the last monitoring value and the first monitoring value of the segment is taken as the growth amplitude (the growth amplitude may be negative, indicating a decrease), and the growth rate is calculated using the growth amplitude (known technology), and the segment with a growth rate greater than a threshold such as 0.5 is taken as an abnormal growth segment, i.e., a segment with a sudden increase in process parameters; In a preferred embodiment of the present application, in each monitoring curve, the product of the mean value of the growth amplitude of the abnormal growth segment and the duration proportion of all abnormal growth segments in the monitoring curve is taken as the fluctuation parameter; the normalized result of the cumulative sum of the fluctuation parameters of all monitoring curves is taken as the process fluctuation coefficient; specifically, the cumulative sum can be mapped into a tanh function to adjust the value range for normalization to obtain the process fluctuation coefficient, or other normalization methods can be used.

[0049] After obtaining the process fluctuation coefficient of each blow molding product in the blow molding production process at the current time, the batch quantity of each batch uneven wall thickness area corresponding to the blow molding product can be further used to preliminarily evaluate whether the batch uneven wall thickness area of the blow molding product is affected by process interference or equipment failure, and further combined with the gray distribution and area change characteristics in the uneven wall thickness area to evaluate the possibility of equipment failure, and then to perform fault warning on the blow molding equipment.

[0050] Preferably, in one embodiment of the present application, considering that the wall thickness unevenness caused by process fluctuations is usually in small batches, the process parameter fluctuations of the corresponding blow molding products are large and the consistency may be extremely low, so the process disturbance coefficient of each batch wall thickness uneven area can be obtained; when the equipment is abnormally malfunctioning, the wall thickness uneven area of the same batch of blow molding products may show an increasingly serious batch trend, which is specifically manifested in that the area and uneven characteristic parameters of the wall thickness uneven area may be stable or increase, so that the fault disturbance coefficient of each batch wall thickness uneven area can be determined; then the batch area affected by the fault can be determined by comprehensively determining the two; then the distribution of the batch area affected by the fault in each blow molding product is analyzed, and the fault influence characteristics of the blow molding product are comprehensively evaluated, and then the equipment fault condition is comprehensively evaluated to give an early warning; Based on this, the method for early warning of blow molding equipment failure comprises: The batch quantity of the corresponding blow molding product of each batch wall thickness uneven area is negatively correlated and mapped to obtain a process disturbance weight; the discrete characteristic value of the process fluctuation coefficient of the corresponding blow molding product is weighted and normalized by using the process disturbance weight to obtain the process disturbance coefficient of each batch wall thickness uneven area; According to the batch accumulation of the area and uneven characteristic parameters of each batch wall thickness uneven area, the fault disturbance coefficient of each batch wall thickness uneven area is obtained, and the process disturbance coefficient is combined to determine the suspected fault affected area from all kinds of batch wall thickness uneven areas; According to the distribution characteristics of the fault disturbance coefficient of the suspected fault affected area in each blow molding product, the production abnormality coefficient of the blow molding product is obtained in combination with the total number and total area of the fault affected area; until the current time, when the production abnormality coefficient of a preset number of blow molding products is greater than a preset early warning threshold value, the blow molding equipment is warned to be malfunctioning.

[0051] As an example, the batch quantity of the corresponding blow molding product of each batch wall thickness uneven area is inversely correlated and mapped, the reciprocal process disturbance weight is obtained, the discrete characteristic value is represented by variance, the process disturbance weight and the variance of the process fluctuation coefficient of the corresponding blow molding product are multiplied, and the product is linearly normalized to obtain the process disturbance coefficient of each batch wall thickness uneven area; In one preferred embodiment of the present application, considering that the sequence is constructed according to the detection order of the blow molding product, the area change and uneven characteristic parameter change of the batch wall thickness uneven area can be more clearly observed, so that the batch accumulation can be evaluated; the fault disturbance coefficient acquisition method comprises: For each type of batch wall thickness uneven area, the area and uneven characteristic parameters are respectively based on the detection sequence of the blow molding product to construct a corresponding change sequence, and according to the number of non-negative values and the cumulative characteristics in the first-order difference sequence of the change sequence, the area growth parameter and the characteristic growth parameter are respectively obtained; the area growth parameter and the characteristic growth parameter are fused to obtain the fault interference coefficient; Specifically, taking the area of each type of batch wall thickness uneven area as an example, the area is sorted according to the wall thickness detection sequence of the corresponding blow molding product to construct an area change sequence, then the area change sequence is first-order differentiated, the number of non-negative difference values is multiplied by the cumulative sum to obtain the area growth parameter; the more non-negative difference values and the larger the cumulative value are, the more stable or growing the wall thickness uneven characteristic is, and the more likely it is to be a batch problem caused by long-term accumulation of equipment failure; similarly, the characteristic growth parameter of the uneven characteristic parameter can be obtained; the area growth parameter and the characteristic growth parameter are multiplied and fused, and the product is linearly normalized to obtain the fault interference coefficient; It is also considered that process fluctuations and equipment failures may occur simultaneously during blow molding, and the process interference coefficient and the fault interference coefficient reflect the possibility of affecting the wall thickness unevenness from different angles; when the process interference coefficient is high and the fault interference coefficient is low, the batch wall thickness uneven area is more likely to be caused by process fluctuations; Based on this, in one preferred embodiment of the present application, the batch wall thickness uneven area with a fault interference coefficient less than or equal to a preset fault threshold and a process interference coefficient greater than a preset process threshold is regarded as a suspected process affected area; the remaining batch wall thickness uneven areas are regarded as suspected fault affected areas; wherein the preset fault threshold is set to 0.3 and the preset process threshold is set to 0.6, so that the suspected process affected areas and the suspected fault affected areas of all batches can be determined; wherein the part of the wall thickness uneven areas without matching results mentioned in step S2 are also suspected process affected areas; After obtaining the suspected fault affected area of the batch of blow molding products, it can be determined whether there is a suspected fault affected area in each blow molding product, and further according to the distribution characteristics of the fault interference coefficient of the suspected fault affected area in each blow molding product, combined with the total number and total area of the fault affected area, the production abnormality coefficient of each blow molding product is evaluated, which reflects the degree of blow molding product abnormality caused by equipment failure, and prepares for subsequent early warning of blow molding equipment failure; Specifically, in each blow molding product, the total number and total area of the suspected fault affected area are summed, the sum is multiplied by the mean value of the fault interference coefficients of all suspected fault affected areas, and then the product is mapped into a sigmoid function to adjust the value range to obtain the production abnormality coefficient; Further, at the current time, when the production abnormality coefficient of a preset number, such as 10 blow molding products, continuously appears to be greater than a preset early warning threshold, such as 0.6, it is automatically determined and early warned that the blow molding equipment has a fault; the implementer can also adjust the determination condition according to actual application; when the fault is early warned, the relevant staff can be prompted to suspend processing to avoid production loss.

[0052] In other embodiments, the implementer can also formulate an early warning emergency degree based on the production abnormality coefficient, and can also early warn the process fluctuation when the process fluctuation coefficient is greater than the preset threshold.

[0053] Based on the same inventive concept, the application also proposes a fault self-diagnosis system of a full-automatic hollow blow molding equipment, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the steps of the fault self-diagnosis method of the full-automatic hollow blow molding equipment described in steps S1-S4 when executing the computer program.

[0054] In summary, the application first acquires all wall thickness uneven areas in the light transmission image of each blow molding product during wall thickness detection, further performs inter-frame matching on the wall thickness uneven areas in different light transmission images to acquire batch wall thickness uneven areas; then acquires a process fluctuation coefficient according to the fluctuation change of the monitoring curve of each process parameter in the blow molding process of the blow molding product; and finally, according to the batch quantity of the corresponding blow molding product of the batch wall thickness uneven area and the process fluctuation coefficient, and in combination with the gray scale distribution and area change characteristics in the wall thickness uneven area, the blow molding equipment is early warned of a fault. The application analyzes the batch condition and cumulative characteristics of the wall thickness uneven area in the blow molding product, and in combination with the fluctuation change of the process parameter, comprehensively evaluates the possibility of being caused by a device fault, so as to accurately perform fault self-diagnosis and early warning on the blow molding equipment.

[0055] It should be noted that: the above-mentioned sequence of the embodiments of the 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 also possible or can be advantageous.

[0056] Each embodiment in the specification is described in a progressive manner, and the same and similar parts between each embodiment can be referred to each other, and each embodiment mainly describes the differences from other embodiments.

Claims

1. A method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus, characterized by, The method comprises: Until the current time, all wall thickness uneven areas in the light transmission image of each blow molding product in the same batch during wall thickness detection are obtained; the monitoring curve of each process parameter during the blow molding process of each blow molding product is obtained, and the process parameters at least include blow molding pressure and mold locking current; According to the position, area and gray scale distribution change of the wall thickness uneven area in the light transmission image, the wall thickness uneven areas in different light transmission images of the blow molding product are matched between frames, and each frame matching result is taken as a batch wall thickness uneven area; For each blow molding product, the process fluctuation coefficient is obtained according to the fluctuation change of the monitoring curve of each process parameter during the blow molding process; at the current time, according to the batch quantity of the blow molding product corresponding to each batch wall thickness uneven area and the process fluctuation coefficient, the gray scale distribution and the area change characteristics in the wall thickness uneven area are combined to perform fault early warning on the blow molding equipment.

2. A method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 1, characterized in that, The method for obtaining the wall thickness uneven area comprises: The continuous light transmission frame and infrared frame of the blow molding product during the wall thickness detection conveying process are obtained, and the light transmission image is included in the continuous light transmission frame; the suspected wall thickness uneven area and uneven characteristic parameters thereof are determined according to the local gray scale distribution in each light transmission frame, and the temperature information of each suspected wall thickness uneven area is obtained by combining the pixel value in the infrared frame with the same frame number; For each blow molding product, the suspected wall thickness uneven area is matched between frames according to the position, area and temperature information of the suspected wall thickness uneven area in the light transmission frame, and the wall thickness uneven area in the corresponding light transmission frame is determined based on the matching result.

3. A method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 2, characterized in that, The method for obtaining the suspected wall thickness uneven area and uneven characteristic parameters thereof comprises: In each light transmission frame, all target points are screened out according to the gradient amplitude of the pixel points, the region connected domain detection is performed on the target points, and the uneven characteristic parameters of each connected domain are obtained according to the gray scale range and the gradient amplitude distribution characteristics of the pixel points in each connected domain; all suspected wall thickness uneven areas are screened out from all connected domains according to the uneven characteristic parameters.

4. The method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 2, characterized in that, The method for obtaining the batch wall thickness uneven area comprises: Any light transmission image is taken as a target image, and any wall thickness uneven area in the target image is taken as a target area; each wall thickness uneven area in each non-target image is taken as a reference area; The first matching parameter is obtained by fusing the center position difference and the area difference between the target area and each reference area, and the negative correlation mapping result of the difference between the uneven characteristic parameters of the corresponding areas is taken as the second matching parameter; the matching coefficient between each reference area and the target area is obtained by fusing the first matching parameter and the second matching parameter; In each non-target image, the reference area between the target area is screened out, which has the maximum matching coefficient and is greater than the preset matching threshold, as the matching area of the target area; the target area and all matching areas thereof are taken as a batch wall thickness uneven area.

5. The method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 1, characterized in that, The method for obtaining the process fluctuation coefficient comprises: For each blow molding product, the monitoring curve of each process parameter is segmented by extreme points, and all abnormal growth sub-sections are screened from all sections according to the growth amplitude of each section; and a process fluctuation coefficient is obtained according to the growth amplitude and duration of all abnormal growth sub-sections in each monitoring curve.

6. A method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 5, characterized in that, The method for obtaining the process fluctuation coefficient according to the growth amplitude and duration of all abnormal growth sub-sections in each monitoring curve comprises: In each monitoring curve, the product of the mean value of the growth amplitude of the abnormal growth sub-section and the duration ratio of all abnormal growth sub-sections in the monitoring curve is taken as a fluctuation parameter; and the normalized result of the cumulative sum of the fluctuation parameters of all monitoring curves is taken as the process fluctuation coefficient.

7. A method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 2, characterized in that, The method for prewarning the blow molding equipment failure comprises: The batch quantity of the blow molding product corresponding to each batch uneven wall thickness region is negatively correlated to obtain a process disturbance weight; and the discrete eigenvalue of the process fluctuation coefficient of the blow molding product is weighted and normalized by the process disturbance weight to obtain the process disturbance coefficient of each batch uneven wall thickness region; According to the area of each batch uneven wall thickness region and the batch accumulation of the uneven characteristic parameter, the failure disturbance coefficient of each batch uneven wall thickness region is obtained, and the suspected failure influence region is determined from all kinds of batch uneven wall thickness regions in combination with the process disturbance coefficient; According to the distribution characteristics of the failure disturbance coefficient of the suspected failure influence region in each blow molding product, in combination with the total quantity and total area of the failure influence region, the production abnormality coefficient of the blow molding product is obtained; and the blow molding equipment failure is prewarned when the production abnormality coefficient of a preset quantity of blow molding products is greater than a preset prewarning threshold value.

8. A method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 7, characterized in that, The method for obtaining the failure disturbance coefficient comprises: For each batch uneven wall thickness region, the area and the uneven characteristic parameter are respectively constructed into corresponding change sequences based on the detection sequence of the blow molding product, and the area growth parameter and the characteristic growth parameter are respectively obtained according to the number and accumulation characteristics of non-negative values in the first-order difference sequence of the change sequence; and the failure disturbance coefficient is obtained by fusing the area growth parameter and the characteristic growth parameter.

9. A method for self-diagnosis of faults in a fully automatic hollow blow molding apparatus according to claim 7, characterized in that, The method for obtaining the suspected failure influence region comprises: The batch uneven wall thickness region with the failure disturbance coefficient less than or equal to a preset failure threshold value and the process disturbance coefficient greater than a preset process threshold value is taken as a suspected process influence region; and the remaining batch uneven wall thickness regions are taken as suspected failure influence regions.

10. A fault self-diagnosis system for a fully automatic hollow blow molding apparatus, characterized by, A computer program product 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 the fault self-diagnosis method of the full-automatic hollow blow molding equipment according to any one of claims 1-9 when executing the computer program.

Images