Flexible film production equipment operation state monitoring system based on big data

Abstract

The invention belongs to the technical field of data processing, and particularly relates to a flexible film production equipment operation state monitoring system based on big data, and the system comprises a data obtaining module which is used for obtaining an operation data sequence of a part and a corresponding production data sequence and environment data sequence; the component life change model and life index construction module is used for acquiring the life change model and life index of each component; the running state model building module is used for acquiring the stability of each related sequence of the part at each moment, acquiring the linear relation between all related sequences of any type of part and the type of part at each moment, and building a running state model of each type of part; and the operation state monitoring module is used for completing state monitoring of each operation part of the flexible film production equipment. According to the invention, the predictability and interpretability of the operation state monitoring of the flexible film production equipment are improved, and the production efficiency of the flexible film is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing. More specifically, the present invention relates to a monitoring system for the operating state of a flexible film production device based on big data. Background Art

[0002] A flexible film is a plastic film composed of multiple materials such as polyethylene and adhesive resin. Due to its transparency, flexibility, water resistance and other characteristics, it is used to make soft packaging, which can conveniently store food and protect products from the external environment during transportation. The production equipment of flexible films includes an extruder, a die runner, a winding device, etc. In order to determine the quality of the flexible film produced in real time and ensure production efficiency, it is often necessary to monitor the operating state of the production equipment.

[0003] In order to determine the operating state of the production equipment, the threshold method is currently often used. The data collected by the equipment sensors is compared with the threshold value to determine whether the operating state of the corresponding equipment and structure is normal. In related technologies, for example, a Chinese patent document with the authorization announcement number CN119373672B discloses a monitoring system for the operating state of a wind power generation equipment, which discloses that by setting a reference deviation threshold value, the operating states of multiple monitoring components are analyzed separately, realizing in-depth analysis of multiple components and comprehensively evaluating the overall health status of the equipment.

[0004] However, the threshold method has an uncertainty, that is, only the real-time state of the equipment can be known. When parts or structures reach the service life or are accidentally damaged, it takes a certain amount of time to debug and maintain the production equipment, which will reduce the production efficiency. On the other hand, the operating state of the flexible film production equipment is affected by various production data and environmental data. Therefore, the threshold method lacks an analysis of the reasons for the operating state of the production equipment and the interpretability of the changes in the operating state of the production equipment, thus affecting the accuracy of detecting the operating state of the production equipment. Summary of the Invention

[0005] To solve the above technical problems of insufficient accuracy and insufficient production efficiency when monitoring the operating state of the production equipment by the threshold method, the present invention provides a monitoring system for the operating state of a flexible film production device based on big data, and the system includes the following modules: A data acquisition module, configured to acquire the operation data sequences of several components of a flexible film production device, acquire several production data sequences and environmental data sequences corresponding to each component, and collectively refer to the production data sequences and environmental data sequences as the relevant sequences of the component; a module for constructing the life change model and life index of the component, configured to obtain the life change model of each component and the life index of each component according to the change characteristics of the operation data sequence of the component over time; an operation state model construction module, configured to obtain the stability of each relevant sequence of the component at each moment according to the change rules of each relevant sequence of the component; according to the stability of each relevant sequence of each component of the same type at each moment and the change correlation of the life indexes of the components of the same type, obtain the linear relationship between all the relevant sequences of any type of component at each moment and the components of the said type; according to the influence degree of all the relevant sequences of the components of the same type at all moments on the components of the same type, establish the operation state models of various components; an operation state monitoring module, configured to complete the state monitoring of each operating component of the flexible film production device according to the operation state models of various components.

[0006] The present invention analyzes the operation data of each type of component, constructs the life change models of various components, enables the full application of the operation data of various components of the flexible film production device, and can more accurately analyze the life change of the flexible film production device; analyzes the influence of production data, environmental data, etc. on the operation data of the component, and thus can avoid the factors that cause the accelerated reduction of the life of the production device during production, and improve the service limit of the device.

[0007] Preferably, the obtaining of the life change model of each component and the life index of each component according to the change characteristics of the operation data sequence of the component over time includes: based on the operation data sequence of the component, obtaining several life characteristics of each component; respectively performing linear fitting on each life characteristic of the target sequence to obtain several life change lines of the target sequence; based on the fitting effect and the line slope of the life change lines of the target sequence, obtaining the life description index of any life change line of the target sequence; taking the life change line with the largest life description index as the life change model of the component corresponding to the target sequence, and taking the life characteristic corresponding to the life change line with the largest life description index as the life index of the target sequence.

[0008] The present invention analyzes various characteristics of different operation data, sets the most effective life indexes for different operation data, and can accurately extract the life changes of different components. Establishing the life change models of different components can improve the accuracy of analyzing and obtaining the life changes of the components during operation.

[0009] Preferably, the obtaining of several life characteristics of each component includes: recording the operation data sequence of any component as the target sequence, obtaining the upper and lower envelope lines of the target sequence, and obtaining the distance between the upper and lower envelope lines of the target sequence at any moment, which is used as the first life characteristic of the target sequence at that moment; obtaining all the data points on the upper and lower envelope lines of the target sequence that intersect with the target sequence, and recording them as the fluctuation characteristic data points of the target sequence, obtaining the distance between any two adjacent fluctuation characteristic data points on the upper envelope line of the target sequence, which is used as the second life characteristic of the target sequence at all moments between the two adjacent fluctuation characteristic data points; taking the data of the target sequence at any moment as the third life characteristic of the target sequence at that moment.

[0010] Preferably, the obtaining of the life description index of any life change line of the target sequence includes: The life description index of the i-th life change line of the target sequence satisfies the expression: ; In the formula, represents the life description index of the i-th life change line of the target sequence; represents the slope of the i-th life change line of the target sequence; represents the set of distances between the points that are fitted to the i-th life change line of the target sequence and the i-th life change line of the target sequence; represents the accumulation function; represents the normalization function.

[0011] The present invention calculates the life description indexes of different life change lines of components, and can establish a life change model of components through the life change line that can most significantly describe the life change of components, and can clearly distinguish different life indexes of components at different usage times.

[0012] Preferably, the obtaining of the stability of each moment of each relevant sequence of the component according to the change rule of the relevant sequence of the component includes: recording any relevant sequence of any component as the target relevant sequence, and recording the c-th moment and the left neighborhood A moment of the target relevant sequence as the first window of the c-th moment of the target relevant sequence; obtaining all the moments before the c-th moment of the target relevant sequence that have the same numerical value as the c-th moment, and recording them as the relevant moments of the c-th moment of the target relevant sequence, and obtaining the first window of the relevant moments of the c-th moment of the target relevant sequence; obtaining the mean value of the Euclidean distances obtained by the coincidence of the first window of the c-th moment of the target relevant sequence and the first windows of all relevant moments, and performing negative correlation normalization, which is recorded as the stability of the c-th moment of the target relevant sequence.

[0013] The present invention analyzes the stability of the relevant sequences of components, so that the influence of the relevant sequences of components on the operation data of components can be based. When the stability of the relevant sequences of components is low and the life index of subsequent operation data changes, it can be obtained that there is a high correlation between the relevant sequences of components and the operation data sequences of components.

[0014] Preferably, obtaining the linear relationship of all relevant sequences of any type of component at each moment with respect to the type of component includes: recording the c-th moment and the right neighborhood A moment of the u-th relevant sequence of any component as the second window of the c-th moment of the u-th relevant sequence of the component; obtaining the correlation between the u-th relevant sequence of any component and the component according to the difference in life index and stability difference of the second window at adjacent moments of the u-th relevant sequence of the component; obtaining the strongly correlated sequences of various components according to the correlation between the u-th relevant sequence of the same type of component and the same type of component; performing multiple linear regression on the numerical values of the strongly correlated sequences of various components at any moment and the life indexes of various components to obtain the linear relationship of all relevant sequences of various components at each moment with respect to the type of component.

[0015] Through multiple linear regression, the present invention can obtain the influence weights of different strongly correlated sequences of components on the change of the life index of components, so as to determine the corresponding change of the life index of components when the strongly correlated sequences of components change.

[0016] Preferably, the correlation between the u-th relevant sequence of any component and the component satisfies the expression: Establish a first threshold function according to the stability difference of adjacent moments of the relevant sequence of the component; ; In the formula, represents the correlation between the u-th relevant sequence of the component and the component; represents the number of numerical values of the operation data sequence of the component; 、 represent the stability of the c-th moment and the c-1-th moment of the u-th relevant sequence of the component; represents the first threshold function; 、 represent the mean values of the life indexes of the second windows of the c-th moment and the c-1-th moment of the operation data sequence of the component; represents the absolute value function; represents the normalization function.

[0017] Preferably, the first threshold function satisfies the expression: ; In the formula, Represents the first threshold function; , Represents the stability of the c-th and (c - 1)-th moments of the u-th relevant sequence of any component; Represents the first threshold.

[0018] Preferably, the obtaining of the strongly relevant sequences of various components includes: taking the mean of the correlations between the u-th relevant sequence of any type of component and all components of that type as the correlation between the u-th relevant sequence of that type of component and that type of component; and denoting the relevant sequences with a correlation greater than the second threshold with respect to that type of component as the strongly relevant sequences of that type of component.

[0019] Preferably, the establishing of the operating state models of various components includes: The operating state model of the z-th type of component is , Represents the set of strongly relevant sequences of the z-th type of component; ; In the formula, Represents the life index of the z-th type of component at the c-th moment; Represents the dependent variable corresponding to the c-th moment of the strongly relevant sequence of the z-th type of component during the multiple linear regression, denoted as the operating state model of the z-th type of component at the c-th moment; Represents the number of values in the operating data sequence of the z-th type of component.

[0020] The beneficial effects of the present invention are as follows: (1) By obtaining the strongly relevant sequences of various components, the present invention screens all relevant sequences, avoiding interference from irrelevant production data and environmental data in the analysis of the operating state of components; (2) The present invention constructs the operating state models of various components, enabling a basis for monitoring the operating state of components during operation, rather than judging the operating state of components by setting thresholds as in the existing methods, and being able to predict the subsequent operating state of components, completing the maintenance of components in advance or preparing the components for subsequent replacement, improving the efficiency of flexible film production; (3) By analyzing the factors affecting the service life of various components, the present invention enables the flexible film production process to avoid these factors, improving the service life of components and reducing the production cost of flexible films. Brief Description of the Drawings

[0021] Figure 1 Is a schematic block diagram showing a system for monitoring the operating state of a flexible film production device based on big data in the present invention; Figure 2 Is a schematic diagram showing the change of the operating data of two components. Detailed implementation manners

[0022] The present invention provides an operation status monitoring system for a flexible film production device based on big data. As Figure 1 shown, an operation status monitoring system for a flexible film production device based on big data includes a data acquisition module 100, a component life change model and a life index construction module 200, an operation status model construction module 300, and an operation status monitoring module 400, which are specifically described below.

[0023] The data acquisition module 100 is used to obtain the operation data sequences of several components of the flexible film production device, and obtain several production data sequences and environmental data sequences corresponding to each component.

[0024] It should be noted that the production device of the flexible film includes components such as an extruder system, a calender roll group, and a winding system. Each component has different monitoring items. For example, the sensors of the winding system include a tension sensor, a laser displacement sensor, and a radial runout detector, which are installed at the guide roller bearing seat, the film edge position, and the winding core shaft end of the winding system respectively. Due to differences in structure and material, the service lives of different positions are different. Therefore, all monitoring items of all components need to be obtained, and each component is analyzed separately.

[0025] It should be noted that during the operation of the winding system, the flexible film is wound into a roll to complete production. The life of the components is worn due to use, and the wear will change due to production-related data and environmental factors. For example, too fast a speed of the winding core shaft will affect the quality of the flexible film and cause accelerated aging of the guide roller bearing seat. If the production environment has a large amount of dust, it will also cause accelerated aging of the production device. Therefore, the present invention obtains production data and environmental data, and then analyzes the influence of production data and environmental data on the device operation data.

[0026] Specifically, obtain the operation data sequences of several components of the flexible film production device, obtain several production data sequences and environmental data sequences corresponding to each component, and collectively refer to the production data sequences and environmental data sequences as the relevant sequences of the corresponding components. There is a corresponding relationship between the operation data sequence, the production data sequence, and the environmental data sequence. The sampling time and sequence length of the operation data sequence of a component are the same as those of the corresponding production data sequence and environmental data sequence. It should be noted that in order to ensure robustness and accuracy, any type of component is not unique, and the obtained components are all components that have reached their service lives.

[0027] So far, the operation data sequences of several components of the flexible film production device have been obtained, and several production data sequences and environmental data sequences corresponding to each component have been obtained.

[0028] The component life change model and life index construction module 200 is used to obtain the life change model and life index of each component according to the change characteristics of the operation data sequence of the component over time.

[0029] It should be noted that taking the guide roll bearing seat of the coiling system as an example, the tension sensor detects the tension of the guide roll bearing seat. When the tension fluctuation is greater than a certain range, it is considered that the guide roll bearing seat has reached its service life. The tension fluctuation range is used as the life index to describe the service life of the guide roll bearing seat. Therefore, it is first necessary to determine the life index of each component.

[0030] It should be noted that the characteristics of the operation data of different components are different, and the changes generated with the wear of the components are also different. For example Figure 2 are the change schematic diagrams of the operation data of two components. Therefore, the life indexes of different components are different. In order to enable the life index to accurately represent the service life of the corresponding component at any time, the life index of each component should be determined in combination with the change characteristics of the operation data over time.

[0031] Specifically, according to the change characteristics of the operation data sequence of the component over time, obtain the life change model and life index of each component: Record the operation data sequence of any component as the target sequence, obtain the upper and lower envelope lines of the target sequence, and obtain the distance between the upper and lower envelope lines at any moment of the target sequence as the first life characteristic of the target sequence at that moment; obtain all the data points on the upper and lower envelope lines of the target sequence that intersect with the target sequence, and record them as the fluctuation characteristic data points of the target sequence. Obtain the distance between any two adjacent fluctuation characteristic data points on the upper envelope line of the target sequence as the second life characteristic of the target sequence at all moments between the two adjacent fluctuation characteristic data points; take the data of the target sequence at any moment as the third life characteristic of the target sequence at that moment.

[0032] Use the least squares method to perform linear fitting on the first life characteristic, second life characteristic, and third life characteristic of all moments of the target sequence to obtain the first life change straight line, second life change straight line, and third life change straight line of the target sequence.

[0033] It should be noted that the first life change straight line, second life change straight line, and third life change straight line of the target sequence characterize the change of the target sequence over time from the fluctuation range, change situation, and change speed of the target sequence. As the life index of the component, it should change significantly over time, rather than changing irregularly or without significant change. Therefore, the better the fitting effect of the first life change straight line, second life change straight line, and third life change straight line of the target sequence, and the greater the degree of change, the more suitable it is as the life change model of the corresponding component of the target sequence.

[0034] The life description exponent of the i-th life change line of the target sequence satisfies the expression: ; In the formula, represents the life description exponent of the i-th life change line of the target sequence; represents the slope of the i-th life change line of the target sequence; represents the set of distances between the points fitted to the i-th life change line of the target sequence and the i-th life change line of the target sequence; represents the accumulation function; represents the normalization function.

[0035] Take the life change line with the largest life description exponent as the life change model of the corresponding component of the target sequence, and take the life characteristics corresponding to the life change line with the largest life description exponent as the life index of the target sequence.

[0036] The operating state model construction module 300 is used to obtain the stability of each relevant sequence of the component at each moment according to the change rules of each relevant sequence of the component; obtain the influence degree of all relevant sequences of any type of component on the component at each moment according to the stability of each relevant sequence of the same type of component at each moment and the change correlation of the life indexes of the same type of component; establish the operating state models of various components according to the influence degree of all relevant sequences of the same type of component on the same type of component at all moments.

[0037] It should be noted that in actual production, the life change model of the component does not show a complete linear decline, but will change due to production data and environmental data. Among all production data sequences and environmental data sequences, there are sequences with a relatively high correlation with the operation data sequence. For example, dust data and the tension fluctuation range of the film winding speed on the guide roller bearing seat have a greater impact, and the equipment operating state is more affected by these factors. Therefore, it is necessary to analyze the influence of all production data sequences and environmental data sequences on each operation data sequence. Moreover, since the influence degree of the relevant sequences that the component can accept at different wear levels is different, the longer the component is used, the smaller the pressure, temperature, etc. that it can bear. Therefore, it is necessary to determine the influence of the relevant sequences of the operation data sequence on the operation data sequence at different moments.

[0038] It should be noted that during the operation of the flexible film production equipment, due to the high requirements for standards such as the thickness of the flexible film, the relevant sequences of the operation data sequence are relatively stable sequences, so as to avoid the flexible film not meeting the production standards. When the relevant sequences of the operation data sequence are stable, the operation data sequence of the component should also change relatively stably according to the life change model. However, during the production process, due to reasons such as personnel movement, mutations in the relevant sequences of the operation data sequence will occur, resulting in the change of the final life index of the operation data sequence not conforming to the life change model. Therefore, the present invention obtains the stable numerical features of each relevant sequence of the operation data sequence, and compares the values of each relevant sequence of the operation data sequence at each moment with the stable numerical features, so as to obtain the stability of each relevant sequence of the operation data sequence at each moment.

[0039] Specifically, according to the change rules of each relevant sequence of the component, the stability of each relevant sequence of the component at each moment is obtained: Denote any relevant sequence of any component as the target relevant sequence, and denote the c-th moment of the target relevant sequence and the A-th moment in the left neighborhood as the first window of the c-th moment of the target relevant sequence. It should be noted that A is set by the implementer according to the actual implementation situation. For example, the value of A can be set to 5.

[0040] Obtain all the moments before the c-th moment of the target relevant sequence with the same value as the c-th moment, and denote them as the relevant moments of the c-th moment of the target relevant sequence. Obtain the first window of the relevant moments of the c-th moment of the target relevant sequence.

[0041] The stability of the c-th moment of the target relevant sequence satisfies the expression: ; In the formula, represents the stability of the c-th moment of the target relevant sequence; represents the number of relevant moments of the c-th moment of the target relevant sequence; represents the subsequence within the first window of the c-th moment of the target relevant sequence, represents the subsequence within the first window of the m-th relevant moment of the c-th moment of the target relevant sequence; represents the Euclidean distance function, which represents the Euclidean distance of the subsequence within the first window after the two windows coincide; represents the exponential function with the natural constant as the base.

[0042] In the formula, It represents the average Euclidean distance obtained after overlapping the first window at the c-th moment of the target-related sequence with the first windows at all related moments. This value indicates the occurrence situation of the c-th moment of the target-related sequence in the existing data. The smaller this value is, the more similar the change situation of the neighborhood of the first window at the c-th moment of the target-related sequence is to that of the same data that has already appeared, and thus the higher the stability of the c-th moment of the target-related sequence.

[0043] Thus, the stability of each moment of each related sequence of the component is obtained.

[0044] It should be noted that when the stability of the target-related sequence is weak and a large change in the life index occurs immediately after the corresponding operation data sequence, it indicates that the change in the life index of the operation data sequence is strongly correlated with the target-related sequence. Thus, the correlation between each component and each related sequence can be obtained. Since the related sequences with strong correlation with the component in each related sequence of the component are not unique, it is impossible to accurately show the specific influence degree of each related sequence on the operation data sequence of the component. And the life indexes and the data of the related sequences of the same type of components at the same moment are not completely the same. Therefore, all the operation data sequences of the same type of components can be synthesized to establish a relational expression between the data of the related sequence at one moment and the life index of the operation data.

[0045] Preferably, according to the stability of each moment of each related sequence of the same type of components and the correlation of the change of the life index of the same type of components, obtain the linear relationship between all related sequences of any type of component at each moment and the type of component: The c-th moment of the u-th related sequence of any component and the right neighborhood A moment are denoted as the second window of the c-th moment of the u-th related sequence of the component. It should be noted that A is set by the implementer according to the actual implementation situation. For example, the value of A can be set to 5.

[0046] The correlation between the u-th related sequence of any component and the component satisfies the expression: ; ; In the formula, represents the correlation between the u-th related sequence of any component and the component; represents the number of values of the operation data sequence of the component; , represent the stability of the c-th moment and the c - 1-th moment of the u-th related sequence of the component; represents the first threshold function. When is greater than the preset first threshold , , when is less than or equal to the preset first threshold When ; , represent the mean values of the life indicators of the second window at the c-th moment and the (c - 1)-th moment of the operation data sequence of the said component; represents the absolute value function; represents the normalization function. It should be noted that the preset first threshold is set by the implementer according to the actual implementation situation. For example the value can be set to 0.1.

[0047] In the formula, represents restricting the change in the stability of adjacent moments of the u-th correlation sequence of the component. When the stability change is large, it is more possible to obtain the correlation between the correlation sequence of the component and the component through the stability change of the correlation sequence and the change in the life indicator of the operation data sequence; represents multiplying and averaging the stability change of the adjacent moments of the moments with large stability changes by the change in the life indicator of the operation data sequence at the corresponding moments, indicating the change consistency between the correlation sequence and the operation data sequence. When the change in the correlation sequence is larger and the change in the operation data sequence is larger, the change consistency between the correlation sequence and the operation data sequence is stronger, thereby indicating that the u-th correlation sequence of the said component is more strongly correlated with the said component.

[0048] Take the mean value of the correlation between the u-th correlation sequence of any type of component and all components of the said type as the correlation between the u-th correlation sequence of the said type of component and the said type of component; Mark the correlation sequences with a correlation greater than the second threshold with the said type of component as the strong correlation sequences of the said type of component. It should be noted that the second threshold is set by the implementer according to the actual implementation situation. For example, the second threshold can be set to 0.5.

[0049] Mark any type of component as the target type component. Take the life indicators of all components of the target category component at the c-th moment as the dependent variable, and take the values of all strong correlation sequences of the target category component at the c-th moment as the independent variable. Use the least squares method for multiple linear regression to obtain the linear relationship between all correlation sequences of each type of component at each moment and the said type of component. The model expression used for the multiple linear regression is , is the feature weight, is the bias, x is the independent variable, is the dependent variable.

[0050] It should be noted that the multiple linear regression models of the target type component at all moments constitute the operation state model of the target type component.

[0051] Preferably, establish the operation state models of each type of component: The operation state model of the z-th type of component is , represents the set of strongly correlated sequences of the z-th type of component; ; In the formula, represents the life index of the z-th type of component at the c-th moment; represents the dependent variable corresponding to the c-th moment of the strongly correlated sequence of the z-th type of component in the process of the multiple linear regression, denoted as the operation state model of the z-th type of component at the c-th moment; represents the number of numerical values in the operation data sequence of the z-th type of component.

[0052] So far, the operation state models of various components have been obtained.

[0053] The operation state monitoring module 400 is used to complete the state monitoring of each operating component of the flexible film production equipment according to the operation state models of various components.

[0054] Specifically, for each component in operation, the numerical value of the strongly correlated sequence of each component at the current moment is input into the corresponding operation state model of each component to obtain the life index of each component, that is, the operation state of each component.

[0055] So far, the operation states of each component of the flexible film production equipment have been obtained.

[0056] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

Claims

1. A monitoring system for the operating state of a flexible film production device based on big data, characterized in that, Including: A data acquisition module, configured to acquire the operation data sequences of several components of a flexible film production device, acquire several production data sequences and environmental data sequences corresponding to each component, and collectively refer to the production data sequences and environmental data sequences as the relevant sequences of the components; A component life change model and life index construction module, configured to obtain the life change models of each component and the life indexes of each component according to the change characteristics of the operation data sequences of the components over time; An operation state model construction module, configured to obtain the stability of each relevant sequence of each component at each moment according to the change rules of each relevant sequence of the component; obtain the linear relationship between all relevant sequences of any type of component at each moment and the component of this type according to the stability of each relevant sequence of each component of the same type at each moment and the change correlation of the life indexes of the components of the same type; establish the operation state models of each type of component according to the influence degree of all relevant sequences of the components of the same type at all moments on the components of the same type; An operation state monitoring module, configured to complete the state monitoring of each operating component of the flexible film production device according to the operation state models of each type of component.

2. The operation status monitoring system of a flexible film production device based on big data according to claim 1, wherein The obtaining of the life change models of each component and the life indexes of each component according to the change characteristics of the operation data sequences of the components over time includes: Based on the operation data sequences of the components, obtaining several life characteristics of each component; respectively performing linear fitting on each life characteristic of the target sequence to obtain several life change lines of the target sequence; obtaining the life description index of any life change line of the target sequence based on the fitting effect and the line slope of the life change lines of the target sequence; Taking the life change line with the largest life description index as the life change model of the component corresponding to the target sequence, and taking the life characteristic corresponding to the life change line with the largest life description index as the life index of the target sequence.

3. The operation status monitoring system of a flexible film production device based on big data according to claim 2, characterized in that, The obtaining of several life characteristics of each component includes: Denoting the operation data sequence of any component as the target sequence, obtaining the upper and lower envelope lines of the target sequence, and taking the distance between the upper and lower envelope lines of the target sequence at any moment as the first life characteristic of the target sequence at this moment; obtaining all the data points on the upper and lower envelope lines of the target sequence that intersect with the target sequence, and denoting them as the fluctuation characteristic data points of the target sequence, and taking the distance between any two adjacent fluctuation characteristic data points on the upper envelope line of the target sequence as the second life characteristic of the target sequence at all moments between the two adjacent fluctuation characteristic data points; taking the data of the target sequence at any moment as the third life characteristic of the target sequence at this moment.

4. A monitoring system for the operating state of a flexible film production device based on big data according to claim 2, characterized in that, The obtaining of the life description index of any life change line of the target sequence includes: The life description index of the i-th life change line of the target sequence satisfies the expression: ； Wherein, represents the life description index of the i-th life change line of the target sequence; represents the slope of the i-th life change line of the target sequence; represents the set of distances between the points that are fitted to the i-th life change line of the target sequence and the i-th life change line of the target sequence; represents the accumulation function; represents the normalization function.

5. A monitoring system for the operating state of a flexible film production device based on big data according to claim 1, characterized in that, The obtaining of the stability of each relevant sequence of each component at each moment according to the change rules of the relevant sequences of the component includes: Denote the arbitrary relevant sequence of any component as the target relevant sequence, and denote the c-th moment and the left neighborhood A moment of the target relevant sequence as the first window of the c-th moment of the target relevant sequence; obtain all the moments before the c-th moment of the target relevant sequence that have the same value as the c-th moment, and denote them as the relevant moments of the c-th moment of the target relevant sequence, and obtain the first window of the relevant moments of the c-th moment of the target relevant sequence; obtain the mean value of the Euclidean distances obtained by the coincidence of the first window of the c-th moment of the target relevant sequence and the first windows of all relevant moments, and perform negative correlation normalization, and denote it as the stability of the c-th moment of the target relevant sequence.

6. The operation status monitoring system for a flexible film production device based on big data according to claim 1, characterized in that, The obtaining of the linear relationship of all relevant sequences of any type of component at each moment with respect to the type of component includes: Denote the c-th moment and the right neighborhood A moment of the u-th relevant sequence of any component as the second window of the c-th moment of the u-th relevant sequence of the component; obtain the correlation between the u-th relevant sequence of any component and the component according to the difference in life index and stability of the second windows of adjacent moments of the u-th relevant sequence of the component; obtain the strongly relevant sequences of all types of components according to the correlation between the u-th relevant sequence of the same type of components and the same type of components; perform multiple linear regression on the values of the strongly relevant sequences of all types of components at any moment and the life indexes of all types of components to obtain the linear relationship of all relevant sequences of all types of components at each moment with respect to the type of component.

7. A monitoring system for the operating state of a flexible film production device based on big data according to claim 6, characterized in that, The correlation between the u-th relevant sequence of any component and the component satisfies the expression: Establish a first threshold function according to the stability difference of adjacent moments of the relevant sequence of the component. ； Wherein, represents the correlation between the u-th relevant sequence of the component and the component; represents the numerical quantity of the operation data sequence of the component; , represents the stability of the c-th and (c - 1)-th moments of the u-th relevant sequence of the component; represents the first threshold function; , represents the mean value of the life indexes of the second window at the c-th and (c - 1)-th moments of the operation data sequence of the component; represents the absolute value function; represents the normalization function.

8. A monitoring system for the operating state of a flexible film production device based on big data according to claim 7, characterized in that, The first threshold function satisfies the expression: ； In the formula, represents the first threshold function; , represent the stability at the c-th and (c - 1)-th moments of the u-th relevant sequence of any component; represents the first threshold.

9. A monitoring system for the operating state of a flexible film production device based on big data according to claim 6, characterized in that, The obtaining of the strongly relevant sequences of all types of components includes: Take the mean value of the correlation between the u-th relevant sequence of any type of component and all components of the type as the correlation between the u-th relevant sequence of the type of component and the type of component; denote the relevant sequences with a correlation greater than the second threshold with respect to the type of component as the strongly relevant sequences of the type of component.

10. A monitoring system for the operating state of a flexible film production device based on big data according to claim 6, characterized in that, The establishment of the operating state model of all types of components includes: The operating state model of the z-th type of component is , which represents the set of strongly correlated sequences of the z-th type of component; ； In the formula, represents the life index of the z-th type of component at the c-th moment; represents the dependent variable corresponding to the c-th moment of the strongly correlated sequence of the z-th type of component in the process of the multiple linear regression, denoted as the operation state model of the z-th type of component at the c-th moment; represents the number of numerical values in the operation data sequence of the z-th type of component.

Images