Intelligent monitoring method and system for brightening composite film piece process based on multi-process parameter fusion

Abstract

This invention relates to the field of material monitoring technology, and more particularly to an intelligent monitoring method and system for the manufacturing process of brightness enhancement composite films based on the fusion of multi-process parameters. The method includes: establishing a unified material segment processing record based on the material segment position record, tension record, and brightness uniformity observation record of the brightness enhancement composite film; reading the reference tension and reference brightness uniformity, calculating the tension coupling coefficient based on the reference tension and reference brightness uniformity, and calculating the material segment state of the brightness enhancement composite film; arranging the material segment states into a one-dimensional sequence, reading the material segment states from the one-dimensional sequence, and calculating the intermediate complete state and the final complete state of the material segment; reading the low treatment threshold and high treatment threshold from a preset database, and calculating the treatment judgment value corresponding to the current material segment in conjunction with the complete state of the material segment; scanning sequentially using the segment number as an index; constructing treatment intervals and treatment boundaries based on the scanning results; and intelligently monitoring the treatment intervals.

Description

Technical Field

[0001] This invention relates to the field of material monitoring technology, and more particularly to a method and system for intelligent monitoring of the brightening composite film manufacturing process based on the fusion of multiple process parameters. Background Technology

[0002] Brightness-enhancing composite films are widely used in display, lighting, and optical module fields. Their production is typically completed using a continuous conveyor belt method, where the film material undergoes multiple processes such as unwinding, tension conveying, lamination, shaping, and online inspection. The final brightness performance, brightness uniformity, and local appearance are not determined by a single process but are gradually formed during continuous processing. Existing production monitoring solutions mostly revolve around equipment-side parameters or single-process inspection results, such as recording tension changes separately, collecting online inspection results separately, and then judging the production status through threshold alarms, final inspection sampling, or single-point anomaly identification. This type of method is useful in judging whether the equipment is operating stably, but it still has significant shortcomings in the scenario of continuous roll material and multi-process cumulative forming of brightness-enhancing composite films. First, the recording position, sampling cycle, and time reference of different processes are not consistent. The tension record of the previous process and the optical inspection result of the subsequent process are often difficult to stably correspond to the same piece of film material. As a result, although the process data is continuously accumulated, it is more at the level of equipment time or inspection time, and it is difficult to directly serve the quality analysis of the same material object. Secondly, anomalies in brightening composite films typically exhibit continuous distribution and delayed manifestation. Deviations in preceding stress levels often do not immediately manifest as defects, but rather gradually appear as decreased brightness uniformity, localized darkening or brightening, or continuous abnormal bands as the film continues its journey to subsequent inspection positions. Therefore, relying solely on single-segment inspection results makes it difficult to accurately determine whether an anomaly has formed, its extent of impact, and appropriate handling. Furthermore, existing solutions primarily rely on single-point assessments, lacking a mechanism to construct a state around continuous material segments and formulate release, re-inspection, or isolation actions accordingly. This can easily lead to problems such as untimely identification of abnormal segments, inaccurate segmentation of continuous abnormal intervals, and reliance on experience for on-site handling. Based on these issues, a monitoring method closer to the actual continuous roll-to-roll process is needed. This method should ensure that key process information and key quality observations are consistently attributed to the same material segment, and on this basis, form a state expression and handling basis oriented towards continuous material segments, thereby improving the accuracy and on-site feasibility of brightening composite film process monitoring. Summary of the Invention

[0003] To address the aforementioned issues, this invention provides a method and system for intelligent monitoring of the brightening composite film manufacturing process based on the fusion of multiple process parameters.

[0004] To achieve the above objectives, this invention proposes, in one aspect, an intelligent monitoring method for the manufacturing process of brightness-enhancing composite films based on the fusion of multiple process parameters, characterized by comprising:

[0005] Establish a unified material segment processing record based on the material segment position record, tension record, and brightness uniformity observation record of the brightness enhancement composite film;

[0006] The reference tension and reference brightness uniformity are read from the process formula, and the tension coupling coefficient is calculated based on the reference tension and reference brightness uniformity. The material segment state of the brightness enhancement composite film is calculated by combining the material segment processing record and the tension coupling coefficient.

[0007] The states of the material segments are arranged into a one-dimensional sequence. The states of the material segments are read from the one-dimensional sequence, and the intermediate complete states of the material segments are calculated. A local deviation enhancement coefficient is introduced, and the final complete state of the material segments is calculated in combination with the intermediate complete states.

[0008] The system reads low and high disposal thresholds from a preset database, calculates the disposal judgment value corresponding to the current material segment based on the complete state of the material segment, scans sequentially using the segment number as an index, constructs disposal intervals and disposal boundaries based on the scan results, visualizes the disposal intervals and disposal boundaries, and performs intelligent monitoring of the disposal intervals.

[0009] In some embodiments, the step of sequentially scanning using the segment number of the material segment as an index, and constructing the treatment interval and treatment boundary based on the scanning results, specifically includes:

[0010] Scan sequentially forward using the segment number of the material segment as the index;

[0011] If the current material segment's disposal judgment value is greater than the first threshold, the segment is included in the disposal interval and marked as an abnormal interval. The disposal judgment value of the material segment adjacent to the current material segment is then determined.

[0012] If the processing judgment value corresponding to an adjacent material segment is greater than the second threshold, the adjacent material segment will be included in the processing interval until the processing judgment value corresponding to the scanned material segment is less than the second threshold, and it will be used as the processing boundary.

[0013] If the current material segment's processing judgment value is greater than the second threshold and no continuous interval is formed on either side of the surrounding material segment, the current material segment will be designated as a re-inspection segment.

[0014] If the current material segment's processing value is less than the second threshold, the current material segment will be treated as a normal release segment.

[0015] In some embodiments, reading the material segment state from the one-dimensional sequence specifically includes:

[0016] For a given material segment, read the current material segment status and the status of adjacent material segments;

[0017] For intermediate material segments, read the status of adjacent material segments;

[0018] For the first paragraph, the state of the first paragraph and the state of the next paragraph constitute a local background.

[0019] For the final section of material, the final state and the previous state together form a local background.

[0020] In some embodiments, the intermediate complete state is generated by calculating an intermediate function, the parameters of which include the material segment state representing the current segment and its adjacent segments before and after it, the continuous background enhancement coefficient, and the boundary modulation coefficient.

[0021] In some embodiments, the complete state of the material segment is generated by calculating a complete state function, the parameters of which include an intermediate complete state, a local deviation enhancement coefficient, and the local mean of the states of the current segment and its two adjacent material segments.

[0022] In some embodiments, the material segment position record is acquired by an encoder and the entire roll of film is divided into continuous material segments according to a pre-set fixed length window; the tension record is continuously acquired by a tension sensor and converted by a data acquisition card to form a tension sequence arranged according to the sampling time sequence; and the brightness uniformity observation record is obtained by an online optical detection device.

[0023] In some embodiments, the tension coupling coefficient is calculated using the representative value of the material segment tension, the reference tension, and the tension amplification factor.

[0024] In some embodiments, the integrity of the material segment includes single-segment deviation, continuous background, and local variation intensity.

[0025] In some embodiments, the intelligent monitoring of the treatment interval specifically includes:

[0026] For abnormal intervals, output the start segment number and end segment number, and generate an isolation instruction in the rewind record or split mark;

[0027] For the re-inspection section, the system generates a re-inspection identifier at the corresponding section number position;

[0028] For normally released sections, maintain the regular flow.

[0029] To achieve the above objectives, another aspect of the present invention proposes an intelligent monitoring system for the manufacturing process of brightness-enhancing composite films based on the fusion of multiple process parameters, comprising:

[0030] The data unification module is used to establish unified material segment processing records based on the material segment position records, tension records, and brightness uniformity observation records of the brightness enhancement composite film;

[0031] The state calculation module is used to read the reference tension and reference brightness uniformity from the process formula, calculate the tension coupling coefficient based on the reference tension and reference brightness uniformity, and calculate the material segment state of the brightness enhancement composite film by combining the material segment processing record and the tension coupling coefficient.

[0032] The complete state calculation module is used to arrange the states of the material segments into a one-dimensional sequence, read the states of the material segments from the one-dimensional sequence, calculate the intermediate complete states of the material segments, introduce the local deviation enhancement coefficient, and calculate the final complete state of the material segments in combination with the intermediate complete states.

[0033] The monitoring and handling module is used to read low and high handling thresholds from a preset database, calculate the handling judgment value corresponding to the current material segment based on the complete status of the material segment, scan sequentially using the segment number of the material segment as an index, construct handling intervals and handling boundaries based on the scanning results, visualize the handling intervals and handling boundaries, and perform intelligent monitoring of the handling intervals.

[0034] The beneficial effects of this invention are as follows:

[0035] This application first establishes material segments based on the operating position of the continuous membrane material, and collects the key process information and key quality observations corresponding to the material segment to form a processing record for a single material segment; then, combined with process reference information, the tension characterization and brightness uniformity observation on the material segment are transformed into the material segment state, so that the front-end process disturbance and the back-end optical performance can be uniformly expressed in the same state quantity; then, the continuous relationship between adjacent material segments is further used to form a complete state, so that continuous abnormal areas, abnormal boundaries and local high-risk segments can be distinguished at the state level;

[0036] Finally, based on the correspondence between the complete state and the handling threshold, the material segment handling result is generated, and combined with the distribution relationship of continuous material segments, release, re-inspection, or isolation intervals are formed. This technical solution no longer focuses on the scattered monitoring of single-process signals, but connects key process information, quality results, and on-site handling actions on continuous film materials into a complete link. This transforms the monitoring object from equipment records to the material segment itself, the judgment basis from single-point observation to continuous state, and the final output from abstract anomaly prompts to directly executable production line handling results. Therefore, it is more suitable for serving the actual production of continuous roll-to-roll products such as brightness enhancement composite films. Attached Figure Description

[0037] Figure 1 This is a flowchart of the intelligent monitoring method for the brightening composite film manufacturing process based on the fusion of multiple process parameters in a specific embodiment of the present invention;

[0038] Figure 2This is a block diagram of an intelligent monitoring system for the brightening composite film manufacturing process based on the fusion of multiple process parameters, as described in a specific embodiment of the present invention. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] refer to Figure 1 As shown, in one aspect of this application embodiment, a method for intelligent monitoring of the brightening composite film manufacturing process based on multi-process parameter fusion is proposed, including:

[0041] S1: Establish a unified material segment processing record based on the material segment position record, tension record, and brightness uniformity observation record of the brightness enhancement composite film, specifically including:

[0042] S1 first establishes a unified material segment processing record for the same section of the brightness-enhancing composite film. The data collection objects are limited to material segment position records, tension records, and brightness uniformity observation records. The material segment position record is acquired by an encoder installed on the unwinding end, traction roller, or main conveyor roller. The encoder continuously outputs displacement pulses as the film moves. The industrial control computer calculates the film's running position based on the accumulated pulse results and divides the entire roll of film into continuous material segments according to a pre-set fixed length window. The fixed length window is set based on the typical scale of on-site anomalies, ensuring that a single material segment can both cover the actual impact range of local brightness anomalies and maintain sufficient resolution. After the material segment division is completed, the system assigns a unique segment number to each segment of film. Subsequent records entering the tension detection area and optical detection area are all collected around this segment number. In the specific data collection, the cumulative position of the encoder is used as a unified segment reference. Combined with the fixed installation distances of the tension detection area and the optical detection area relative to this reference, the membrane material position corresponding to the sampling time at each workstation is converted to the corresponding material segment. When a sampling window simultaneously covers two adjacent segments, they can be assigned to adjacent material segments according to the proportion of the window coverage length, or, if the coverage proportion is significantly biased to one side, directly assigned to the material segment with the higher proportion. This established material segment position framework ensures that the continuous membrane material always corresponds to the same processing object during multi-process operation, facilitating subsequent processing directly around the material segment.

[0043] Tension records are continuously collected by tension sensors mounted on tension rollers or detection rollers, and converted by a data acquisition card into a tension sequence arranged according to the sampling time order. Since multiple consecutive sample values ​​are generated when the same segment of membrane material passes through the tension detection zone, the system uses the discrete sample mean formula from statistics to convert all tension sample values ​​of the same material segment within the detection zone into a single representative tension value. The application of the classic mean formula here is to explicitly limit the statistical object to all sample values ​​of the same material segment within the tension detection zone, replacing the original continuous sequence with a segment-level representative value, thereby transforming the time-series process record into a material segment-level process characterization quantity. To ensure the stability of the segment-level representative value, tension samples within the same material segment can first eliminate abnormal values ​​such as sensor malfunctions, instantaneous jumps, or values ​​significantly exceeding the process allowable range, and then calculate the average of the remaining valid samples. When the number of valid tension samples in a segment is lower than the preset minimum number of samples, the segment can be marked as a low-confidence segment, and the statistical results of adjacent valid segments can be used for temporary replacement, or the mark can be directly retained for priority processing in subsequent re-inspection stages. The representative tension value is calculated according to the following formula:

[0044] ;

[0045] in, Indicates the first The representative value of the tension of the segment material; Indicates the first When the material segment passes through the tension detection position, the first Tension records obtained from the second sampling; Indicates the first The number of tension sampling times corresponding to the material segment; The segment number is determined by the encoder's cumulative displacement and the preset material segment length. This indicates the sampling sequence number of the tension record within the material segment. Both the left and right sides of the formula represent the same type of tension characterization quantity. When the original tension record is represented by a proportional value, the calculation result is still in proportional form; when the original tension record is represented by a normalized result, the calculation result remains in normalized form.

[0046] Brightness uniformity observation records are obtained by an online optical inspection device, preferably using a line scan camera with a stable light source to continuously scan the membrane surface along the membrane material's running direction. The inspection module then outputs the brightness uniformity observation results corresponding to the currently scanned area. The inspection system assigns the observation results to the corresponding material segment based on the fixed relationship between the inspection window and the material segment's position. For multiple consecutive inspection windows corresponding to the same material segment, a weighted average based on the window coverage length can be used to form a unique brightness uniformity observation value for that segment, or the statistical value can be directly used if the inspection device has already output segment-level statistical values. When an inspection window crosses the boundary between two adjacent segments, it is also aggregated according to its respective coverage ratio, thus ensuring that the tension records and brightness uniformity observation records for the same membrane material segment are included in the same segment-level processing record.

[0047] In one embodiment, a production line sets the material segment length as a fixed window. When the twelfth segment passes through the tension detection zone, it obtains five proportionalized tension sampling values: 0.98, 0.99, 1.01, 1.00, and 1.02. Substituting these values ​​into the formula above, the representative tension value for this segment is 1.00. Subsequently, the twelfth segment continues to the online optical detection position. The detection module outputs the corresponding brightness uniformity observation value of 0.93. Based on this, the system generates a processing record for the twelfth segment, including the segment number, the representative tension value of 1.00, and the brightness uniformity observation value of 0.93. Through this aggregation, key records originally scattered across different processes and times are integrated into a processing record for a single material segment. Subsequent steps can directly generate the material segment status based on this record without needing to perform cross-process matching again. This material segment processing record directly serves the core objective of multi-process parameter fusion for the brightness enhancement composite film, stably mapping the stress process and final brightness uniformity performance, which are of most concern on-site, to the same segment of film material.

[0048] S2: Read the reference tension and reference brightness uniformity from the process formula, calculate the tension coupling coefficient based on the reference tension and reference brightness uniformity, and calculate the material segment state of the brightness enhancement composite film by combining the material segment processing record and the tension coupling coefficient, specifically including:

[0049] In S2, the material segment processing record generated in S1 is directly used as input to generate the corresponding material segment state for each brightening composite film segment. S1 has already stably attributed the tension representative value and brightness uniformity observation results of the same film segment across multiple processes to the same material segment processing record. Therefore, this step no longer handles cross-process alignment issues, but instead focuses on calculating "what quality formation state this film segment is currently in." In actual implementation, the system reads the segment number, tension representative value, and brightness uniformity observation results from each material segment processing record, and simultaneously reads the reference tension and reference brightness uniformity from the corresponding product specification's process formula. The reference tension originates from the historical qualified segment statistics during stable mass production. Specifically, under the same product specification, the same production line configuration, and the same process window, multiple qualified roll material segment processing records are continuously extracted, and their tension representative values ​​are taken at the center level. The reference brightness uniformity originates from the statistical results of qualified segments in the same batch during online optical inspection, and the value is determined in the same way as the reference tension. The input data thus constructed all have clear sources and correspond to the same film segment. Subsequent state generation is performed segment by segment based on this set of segment-level data. For the same product specifications and during the same roll production process, , , and It can remain fixed, only being updated after changing specifications, formulas, or completing recalibration, to ensure that the results of the same batch of material segments can be directly compared under the same evaluation criteria. The actual process characteristics of brightness enhancement composite films are that pre-process tension deviation has a continuous impact on subsequent optical performance. This impact is relatively gradual when approaching the process baseline, but gradually amplifies as the deviation increases. Therefore, this step first uses the classical quadratic deviation term to construct a tension coupling coefficient, and then uses this coupling coefficient to correct the brightness deviation, ultimately obtaining the material segment state. The original source of the quadratic deviation term is the amplified expression of a quadratic function in mathematics, often used to describe the non-linear increase in the degree of influence as the deviation magnitude increases. The specific modification made in this application is to limit the quadratic term, which originally only described the abstract magnitude of the deviation, to the square of the deviation of the representative tension value of the same material segment relative to the reference tension, and to introduce a calibrable process amplification coefficient before it, so that it directly serves the generation of the segment-level quality state of the brightness enhancement composite film. Correspondingly written as:

[0050] ;

[0051] in, Indicates the first The tension coupling coefficient of the segment material; Indicates the first The tension representative value of the segment material is derived from the tension representative value field in the processing record of the S1 material segment. This represents the reference tension for the current product specification, which is derived from the reference database corresponding to the process formula. The tension amplification factor, derived from calibration results during the trial production phase, is obtained by selecting historical material processing records of qualified and abnormal sections and comparing the correlation between tension deviation and subsequent brightness anomalies. Both the left and right sides of the formula represent the number of coupling systems of the same type, and proportionalization maintains the same evaluation scale. The squared term maps both positive and negative deviations to deviation magnitudes, consistent with the common process principle that excessive or insufficient stress on the membrane material can affect subsequent brightness uniformity. The constant term indicates that when the tension equals the reference tension, the coupling factor remains at the baseline level. This yields... This can be understood as "the degree to which the brightness deviation is amplified under tension background of the current membrane material". It is the direct input for the next state calculation, so there is a clear logical relationship between the formulas.

[0052] After obtaining the tension coupling coefficient, the material segment state is generated according to the classical linear combination approach. The original source of linear combination is the weighted sum expression in mathematics, used to synthesize multiple factors normalized to the same evaluation space into a single state quantity. The specific derivation in this application is to first use the deviation of the representative tension value relative to the reference tension as a preceding process disturbance term, then use the deviation of brightness uniformity relative to the reference brightness uniformity as a current optical performance term, and use the aforementioned tension coupling coefficient to modulate the latter, thereby allowing the material segment state to carry both the preceding stress history and the current brightness enhancement result. This can be written as:

[0053] ;

[0054] in, Indicates the first The material segment state of the segment material; The weight of a single deviation of tension is derived from the calibration results of historical records during the process debugging phase, and is specifically determined by comparing the sensitivity of tension deviation with subsequent brightness deviation. This indicates the baseline brightness uniformity under the current product specifications, derived from the optical statistical results of the qualified production segment; Indicates the first The brightness uniformity observation results of the segment material are derived from the optical observation field in the processing record of the S1 material segment. This represents the tension coupling coefficient calculated by the previous formula. Both the left and right sides of this formula represent state variables of the same type of material segment. Tension deviation and brightness deviation terms are expressed as proportional values ​​or normalized results before being included in the calculation, thus allowing for superposition within the same state scale. The derivation is as follows: first, the coupling coefficient is generated from the tension deviation; then, the brightness deviation is corrected using the coupling coefficient; finally, it is superimposed with the first-order tension deviation term to form the state value. When a material segment lacks an effective representative tension value or an effective brightness uniformity observation value, the final state of that segment can be temporarily not generated, and it can be marked as a segment to be re-inspected. When both adjacent effective segments exist, the statistical results of adjacent effective segments can be used to form temporary supplementary values ​​for continuous calculation, but the supplementary markings should be retained to avoid confusion with the measured state. The state value constructed in this way can reflect the scenario characteristic that "the same magnitude of brightness deviation corresponds to different risk levels under different tension backgrounds," which is consistent with the engineering law that the stress history of the preceding process affects the optical uniformity of the subsequent process in the continuous manufacturing process of brightness enhancement composite films. For example, in one set of embodiments, the reference tension for a certain product specification is 0.98, the reference brightness uniformity is 0.96, the tension amplification factor is 0.8, and the first deviation weight of tension is 0.4. In the material processing record of a certain material segment, the representative tension value is 1.02, and the observed brightness uniformity result is 0.93.

[0055] First, calculate the tension deviation: 1.02 - 0.98 = 0.04. Substituting this into equation 1, the coupling coefficient equals (1 + 0.8)^(0.04)^2. The square of 0.04 is 0.0016, which is then multiplied by 0.8 to get 0.00128. Therefore, the coupling coefficient is 1.00128. Next, calculate the brightness deviation: 0.96 - 0.93 equals 0.03. The first-order tension deviation is 0.4 multiplied by 0.04, resulting in 0.016. The brightness deviation after coupling correction is 1.00128 multiplied by 0.03, resulting in 0.0300384. After adding the two terms, the material segment state value is 0.0460384. Looking at another material segment, if its brightness uniformity is also 0.93, but its tension representative value is 0.99, then the tension deviation is only 0.01. Substituting the same parameters, the coupling coefficient is approximately 1.000008, the tension deviation term is 0.004, the brightness deviation term after coupling correction is approximately 0.0300024, and the final state value is approximately 0.0340024. The brightness observation results of the two material segments are the same. However, the former has a larger tension background deviation, so its state value is higher. This state result can distinguish material segments with "similar current brightness but different formation processes" earlier, which is beneficial for subsequent continuous state completion and handling judgment of material segments. After this step, each material segment processing record generates a corresponding material segment state, which becomes the direct input for the next step to continue forming a complete state.

[0056] S3: Arrange the states of the material segments into a one-dimensional sequence, read the states of the material segments from the one-dimensional sequence, calculate the intermediate complete states of the material segments, introduce a local deviation enhancement coefficient, and calculate the final complete state of the material segments based on the intermediate complete states. Specifically, this includes:

[0057] In this step, the material segment states obtained in S2 are used as continuous inputs to continuously reconstruct the state sequence of the same roll of brightening composite film along the running direction. This ensures that the state of each material segment retains its deviation degree obtained from the coupling of tension and brightness, while also reflecting the continuous influence relationship between adjacent material segments. The material segment states obtained in S2 are already a unified dimensionless state quantity, derived from the superposition of deviations between the representative tension value and the observed brightness uniformity value, and have been normalized to fall on the same evaluation scale. Therefore, in this step, they can be directly used as the basis quantity for subsequent continuous calculations. The unified evaluation scale here means that the representative tension value, reference tension, observed brightness uniformity value, and reference brightness uniformity involved in the calculation in S2 are all input using proportionalized or normalized results. Therefore, the generated material segment states can be directly superimposed and compared between adjacent segments. Continuous membrane materials exhibit significant spatial continuity in actual production, meaning that anomalies in one segment often propagate to the preceding and following segments. Furthermore, acceleration or deceleration of state changes occurs at the boundary of anomalies. Therefore, it is necessary to introduce information from adjacent segments on top of the single-segment state to form a complete state.

[0058] In practice, the material segment states are arranged in order of segment number to form a one-dimensional sequence, and for the first segment... Segment material, read the current segment status and the adjacent preceding state and the next state For intermediate material segments, the states of adjacent segments are directly read; for the first segment, the states of the first segment and the next segment constitute the local background; for the last segment, the states of the last segment and the previous segment constitute the local background. Based on the original idea of ​​local weighted averaging in signal processing, the states of adjacent segments are used to construct local background values. Furthermore, modifications are introduced to address the abnormal distribution characteristics of continuous membrane materials. This involves considering not only simple averaging but also the changing trends between the current segment and adjacent segments, enabling the complete state to simultaneously express the differences between continuous and boundary regions. The local background part uses the averaging of adjacent segment states, which is the most basic form of local estimation in the classic sliding window method. Based on this, a modulation term based on differential product is introduced to enhance the expression of the state at the boundary positions. This differential product originates from the discrete difference method and is used to characterize the direction and magnitude of local changes. The modification in this application is to use the product of two adjacent differences as a boundary strength index, causing it to produce a significant response when the sign or amplitude changes significantly. Based on the above logic, an intermediate function is constructed, and its expression is:

[0059] ;

[0060] in, Indicates the first The intermediate intact state of the segment material; , , These represent the material segment states of the current segment and its adjacent segments before and after it. These states are all derived from the calculation results of S2. For the first and last segments, the adjacent segment states are determined according to the local background composition of the corresponding boundary positions. Specifically, the missing side can be supplemented with the current segment state, or the missing side can be supplemented with the state of the only existing adjacent segment, so that the boundary segment still maintains the same calculation structure as the middle segment. It is preferable to use the current segment state to supplement the missing side. This represents the continuous background enhancement coefficient, the value of which is determined during the production line commissioning phase based on the typical length distribution of continuous abnormal sections. This represents the boundary modulation coefficient, the value of which is determined by the identification requirements of the anomalous boundary segment. The first term of the formula retains the state information of the current segment, the second term introduces the average value of adjacent segments as continuous background enhancement, and the third term uses the absolute value of the product of adjacent differences to characterize the intensity of local changes, thus obtaining different state expressions within the continuous anomalous region and at the anomalous boundary location. Because... All three terms in the series are the same normalized state variables, and they have the same numerical dimensions, so they can be directly weighted and superimposed.

[0061] Obtaining the intermediate complete state Subsequently, to further enhance the ability to identify local abnormal peak segments, a complete state function is introduced, which is based on an enhancement term for local center deviation. This enhancement term originates from the centralized squared deviation expression in statistics, whose original form is used to measure the degree of deviation of an observation value from the mean. In this application, it is limited to the squared deviation of the current segment from its local three-segment mean, multiplied by an enhancement coefficient, thereby highlighting local high-risk segments without changing the overall trend. The complete state function is written as:

[0062] ;

[0063] in, This indicates the final complete state of the material segment; This represents the intermediate complete state obtained from the calculation in the previous formula; This represents the local deviation enhancement coefficient, the value of which is determined by the debugging results of the sensitivity of abnormal segment identification during the trial production stage; This represents the local mean of the current segment and its two adjacent segments. Structurally, this formula forms a continuous derivation relationship with the previous formula. First, continuous background and boundary information are obtained through the product of local averages and differences. Then, the high-deviation segments are enhanced by the squared deviation of the local mean, thus forming a complete state that contains both continuous information and highlights local extrema. When there are states awaiting re-inspection or states temporarily filled in by preceding and following segments in the three local segments, the missing values ​​can be filled in by adjacent valid states before calculating the complete state. The output result retains a mark indicating that the segment was involved in the calculation by the filling, so that it can be prioritized for re-inspection rather than direct release during subsequent processing.

[0064] The following is an illustration using a set of examples. Assume that the three consecutive material states output by S2 are as follows: , , ,Pick , , First, calculate the average value of adjacent segments in the twelfth segment. Adjacent differences are , The absolute value of their product is approximately Substituting into the first equation, we get... Then, the mean of the three local segments was calculated. ,

[0065] Deviation amount is After squaring, it is approximately multiplied by Get about ,and Adding them together yields the final complete state. The same method can be used to calculate other material segments one by one. Through this process, it can be seen that the state of the twelfth segment is further enhanced among the three consecutive segments. This is because it is both in the continuous deviation region and shows a significant deviation from the local mean, thus being strengthened in the complete state.

[0066] Through the above continuous derivation, each material segment ultimately obtains a complete state. This state numerically integrates information from three aspects: single-segment deviation, continuous background, and local variation intensity. It can better reflect the actual quality distribution characteristics of the brightness enhancement composite film in the continuous process. This complete state serves as a direct input for subsequent processing steps, allowing subsequent judgments to simultaneously consider single-segment risks and continuous area risks, thereby improving the accuracy of identifying and handling abnormal segments.

[0067] S4: Read the low and high disposal thresholds from the preset database, and calculate the disposal judgment value corresponding to the current material segment based on the integrity status of the material segment. Scan sequentially using the segment number as the index, construct disposal intervals and disposal boundaries based on the scan results, visualize the disposal intervals and disposal boundaries, and intelligently monitor the disposal intervals, specifically including:

[0068] In S4, the complete state sequence of material segments formed in S3 is directly used as the basis for processing, transforming the continuous state results into material segment processing results that can be executed on the production line. The complete state obtained in S3 has already compressed the single-segment state, the continuous background of adjacent segments, and the intensity of local changes into the same evaluation scale. Therefore, this step no longer returns to the original process record level, but instead focuses on the complete state sequence itself for hierarchical judgment and interval processing. For continuous roll products such as brightening composite films, what is truly of concern on-site is not a particular abstract state value itself, but whether the material segment corresponding to that state value should be released normally, enter re-inspection, or be isolated in the winding or slitting process. In actual production, if the same single-segment deviation occurs in isolation, it usually corresponds to local fluctuations or boundary disturbances, which are more suitable for re-inspection; while if several adjacent material segments continuously maintain a high integrity state, it is closer to a continuous abnormal zone, which is suitable for directly forming an isolation interval. Therefore, based on the single-segment judgment, this step further utilizes the continuous sequence relationship of material segments to form interval processing results, so that the state expression naturally falls into the production line action.

[0069] In practice, the system reads the complete status sequence sequentially according to the material segment number, and then... Section material, read the complete state output by S3 Simultaneously, it reads the low disposal threshold from the process quality database corresponding to the current product specification. and high processing threshold Both thresholds are derived from the historical material integrity distribution of products of the same specification during stable production. The lower threshold distinguishes between normal sections and sections requiring attention, while the higher threshold distinguishes between sections requiring attention and abnormal sections. In practice, this can be achieved using quantile statistics of historical integrity sequence data. For example, the upper edge of the integrity distribution of qualified roll materials can be used as the lower threshold, and the lower edge of the integrity distribution of confirmed abnormal roll materials can be used as the higher threshold. The advantage of this approach is that the thresholds are directly linked to specific product specifications and actual production line conditions, ensuring that subsequent judgments are consistent with existing quality experience on-site. To avoid instability in normalization due to excessively small intervals between the high and low thresholds, when... and When the difference is less than the preset minimum spacing, the current threshold can be paused for direct determination, and the threshold configuration of the most recent stable batch of the same specification can be called, or the determination value can be calculated after expanding according to the preset minimum spacing. Based on the above input, this step first adopts the linear normalization approach to map the complete state to a unified determination value. The original source of the linear normalization method is the interval mapping formula in mathematics, which is often used to convert values ​​in different ranges to a unified comparison scale; the specific modification of this application is to directly set the mapping interval to the low and high treatment thresholds in the process quality database, so that the normalization result is not just a general numerical transformation, but directly expresses the positional relationship of the material segment relative to the treatment boundary. Correspondingly, the first The disposal judgment value for the segment material is written as:

[0070] ;

[0071] in, Indicates the first The disposal judgment value of the segment material is used to generate the disposal category of the segment material; Indicates the first The complete state of the material segment is derived from the calculation results of S3; This indicates a low disposal threshold, derived from the statistical results of the complete status of historical qualified segments for the current product specification. The high disposal threshold is derived from the statistical results of the historical abnormal segment integrity status of the current product specification. Both the left and right sides of the equation represent the normalized disposal judgment quantity. Integrity status and the threshold are on the same state scale; therefore, the numerator and denominator maintain a consistent evaluation dimension, resulting in a dimensionless judgment value that meets the requirements for continuous comparison and threshold grading. The derivation of this equation involves first subtracting the current material segment's integrity status from the low disposal threshold to obtain the degree of deviation of that segment from the boundary of concern. Then, normalization is performed using the interval between high and low thresholds, thus mapping the integrity status of different batches and specifications to a directly interpretable disposal scale. When the value is less than zero, it indicates that the complete state of this segment is below the low processing threshold, corresponding to normal release; when When the value is between zero and one, it indicates that the paragraph is within the scope of interest, and a review and judgment are required; when... When the value is greater than 1, it indicates that the segment exceeds the high handling threshold and meets the conditions for forming an abnormal handling interval.

[0072] After calculating the single-segment handling judgment value, the system continues to scan the interval along the material segment sequence. A continuous interval expansion algorithm is used here, based on the fact that roll material anomalies often exist in the form of continuous segments in actual production. During algorithm implementation, the system scans sequentially using the segment number as an index: when the handling judgment value of a material segment is greater than one, that segment is designated as the core anomaly segment; subsequently, the handling judgment values ​​of adjacent material segments are checked forward and backward. If the handling judgment value of an adjacent segment is greater than zero, it is merged into the same handling interval until a material segment with a handling judgment value less than zero is encountered. The resulting interval directly corresponds to the subsequent segmentation and isolation range. During interval scanning, material segments with a handling judgment value of zero can be considered as boundary segments of concern rather than core anomaly segments; when such a boundary segment is adjacent to a core anomaly segment, it can also be output as an isolation boundary segment to reduce overly fragmented segmentation on-site. For material segments with a handling judgment value greater than zero but without continuous intervals on either side, their handling category is determined as a re-inspection segment; for material segments with a handling judgment value less than zero, their handling category is determined as a normal release segment. The above process relies on the natural sequence of material segments and the continuous distribution of disposal judgment values. The implementation path is clear and easy to directly write into the MES system, production line dashboard, or winding control record. Using a set of calculation examples, let's assume that the complete states of the three consecutive material segments output by S3 are as follows: , , The current product specification corresponds to a low disposal threshold. High processing threshold After substituting into the formula, the decision value for the eleventh paragraph is... The judgment value for the twelfth paragraph is... The judgment value for the thirteenth segment is... Based on the above results, segment 12 is greater than 1 and is identified as the core abnormal segment; segments 11 and 13 are both greater than zero but not greater than 1, indicating that they are in the area of ​​concern. When further expanding the interval, since the adjacent segments on both sides of segment 12 maintain positive values, the system classifies segments 11, 12, and 13 into the same abnormal handling interval. For on-site execution, this means that the winding or slitting process no longer isolates only segment 12, but forms a continuous isolation interval for segments 11 to 13, thus better reflecting the true distribution pattern of abnormal zones in continuous membrane materials. If, in another group of material segments, a segment has a handling judgment value of 0.3, and both its preceding and following segments are less than zero, then the handling category of that segment is determined as a re-inspection segment, which is then carefully confirmed by the operator at the re-inspection station or re-testing process.

[0073] After the above processing, each material segment obtains a single-segment handling judgment value, while continuous abnormal segments obtain interval-based handling boundaries. The system then writes the handling results into the production database and outputs them on the operation interface in the form of material segment number, handling category, and interval start and end positions. For normal release segments, the system maintains normal flow; for re-inspection segments, the system generates a re-inspection mark at the corresponding segment number position and prompts the operator to focus on checking this segment in subsequent quality inspection stages; for abnormal intervals, the system outputs the start segment number and end segment number and generates an isolation instruction in the rewinding record or slitting mark. In this way, the complete state formed in S3 is naturally transformed into specific process action results. The technical effect of the entire process is that the handling result does not depend on the single segment state in isolation, but on the complete state and the continuous relationship of material segments. Therefore, it can simultaneously identify local risk segments and continuous abnormal areas, which is closer to the actual production needs in continuous roll material scenarios such as brightness enhancement composite films. In terms of system implementation, the above method can be composed of a segment position information acquisition unit, a tension and optical data collection unit, a material segment state generation unit, a complete state formation unit, and a disposal output unit, and is uniformly deployed on an industrial control computer or production line monitoring server, thereby forming a brightening composite film process intelligent monitoring system corresponding to the aforementioned method.

[0074] refer to Figure 2 As shown, in another aspect of the embodiments of this application, a brightening composite film manufacturing process intelligent monitoring system based on multi-process parameter fusion is also proposed, including:

[0075] The data unification module is used to establish unified material segment processing records based on the material segment position records, tension records, and brightness uniformity observation records of the brightness enhancement composite film;

[0076] The state calculation module is used to read the reference tension and reference brightness uniformity from the process formula, calculate the tension coupling coefficient based on the reference tension and reference brightness uniformity, and calculate the material segment state of the brightness enhancement composite film by combining the material segment processing record and the tension coupling coefficient.

[0077] The complete state calculation module is used to arrange the states of the material segments into a one-dimensional sequence, read the states of the material segments from the one-dimensional sequence, calculate the intermediate complete states of the material segments, introduce the local deviation enhancement coefficient, and calculate the final complete state of the material segments in combination with the intermediate complete states.

[0078] The monitoring and handling module is used to read low and high handling thresholds from a preset database, calculate the handling judgment value corresponding to the current material segment based on the complete status of the material segment, scan sequentially using the segment number of the material segment as an index, construct handling intervals and handling boundaries based on the scanning results, visualize the handling intervals and handling boundaries, and perform intelligent monitoring of the handling intervals.

[0079] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for intelligent monitoring of the brightening composite film manufacturing process based on multi-process parameter fusion, characterized in that, include: Establish a unified material segment processing record based on the material segment position record, tension record, and brightness uniformity observation record of the brightness enhancement composite film; The reference tension and reference brightness uniformity are read from the process formula, and the tension coupling coefficient is calculated based on the reference tension and reference brightness uniformity. The material segment state of the brightness enhancement composite film is calculated by combining the material segment processing record and the tension coupling coefficient. The states of the material segments are arranged into a one-dimensional sequence. The states of the material segments are read from the one-dimensional sequence, and the intermediate complete states of the material segments are calculated. A local deviation enhancement coefficient is introduced, and the final complete state of the material segments is calculated in combination with the intermediate complete states. The system reads low and high disposal thresholds from a preset database, calculates the disposal judgment value corresponding to the current material segment based on the complete state of the material segment, scans sequentially using the segment number as an index, constructs disposal intervals and disposal boundaries based on the scan results, visualizes the disposal intervals and disposal boundaries, and performs intelligent monitoring of the disposal intervals.

2. The intelligent monitoring method for the manufacturing process of brightness-enhancing composite films based on multi-process parameter fusion according to claim 1, characterized in that, The process of sequentially scanning using the segment number of the material segment as an index, and constructing the treatment interval and treatment boundary based on the scanning results, specifically includes: Scan sequentially forward using the segment number of the material segment as the index; If the current material segment's disposal judgment value is greater than the first threshold, the segment is included in the disposal interval and marked as an abnormal interval. The disposal judgment value of the material segment adjacent to the current material segment is then determined. If the processing judgment value corresponding to an adjacent material segment is greater than the second threshold, the adjacent material segment will be included in the processing interval until the processing judgment value corresponding to the scanned material segment is less than the second threshold, and it will be used as the processing boundary. If the current material segment's processing judgment value is greater than the second threshold and no continuous interval is formed on either side of the surrounding material segment, the current material segment will be designated as a re-inspection segment. If the current material segment's processing value is less than the second threshold, the current material segment will be treated as a normal release segment.

3. The intelligent monitoring method for the brightening composite film manufacturing process based on multi-process parameter fusion according to claim 1, characterized in that, The step of reading the material segment state from the one-dimensional sequence specifically includes: For a given material segment, read the current material segment status and the status of adjacent material segments; For intermediate material segments, read the status of adjacent material segments; For the first paragraph, the state of the first paragraph and the state of the next paragraph constitute a local background. For the final section of material, the final state and the previous state together form a local background.

4. The intelligent monitoring method for brightness enhancement composite film manufacturing process based on multi-process parameter fusion according to claim 1, characterized in that, The intermediate complete state is generated by calculating an intermediate function. The parameters of the intermediate function include the material segment state representing the current segment and its adjacent segments before and after it, the continuous background enhancement coefficient, and the boundary modulation coefficient.

5. The intelligent monitoring method for the brightening composite film manufacturing process based on multi-process parameter fusion according to claim 1, characterized in that, The complete state of the material segment is generated by calculating the complete state function. The parameters of the complete state function include the intermediate complete state, the local deviation enhancement coefficient, and the local mean of the state of the current segment and its two adjacent material segments.

6. The intelligent monitoring method for the manufacturing process of brightness-enhancing composite films based on multi-process parameter fusion according to claim 1, characterized in that, The material segment position record is acquired by the encoder and the entire roll of film is divided into continuous material segments according to a pre-set fixed length window. The tension record is continuously acquired by the tension sensor and converted by the acquisition card to form a tension sequence arranged according to the sampling time sequence. The brightness uniformity observation record is obtained by the online optical detection device.

7. The intelligent monitoring method for the manufacturing process of brightness-enhancing composite films based on multi-process parameter fusion according to claim 1, characterized in that, The tension coupling coefficient is calculated using the representative value of the material segment tension, the reference tension, and the tension amplification factor.

8. The intelligent monitoring method for the manufacturing process of brightness-enhancing composite films based on multi-process parameter fusion according to claim 1, characterized in that, The complete state of the material segment includes single-segment deviation, continuous background, and local variation intensity.

9. The intelligent monitoring method for the manufacturing process of brightness-enhancing composite films based on multi-process parameter fusion according to claim 1, characterized in that, The intelligent monitoring of the treatment area specifically includes: For abnormal intervals, output the start segment number and end segment number, and generate an isolation instruction in the rewind record or split mark; For the re-inspection section, the system generates a re-inspection identifier at the corresponding section number position; For normally released sections, maintain the regular flow.

10. A smart monitoring system for the manufacturing process of brightness-enhancing composite films based on multi-process parameter fusion, characterized in that, include: The data unification module is used to establish unified material segment processing records based on the material segment position records, tension records, and brightness uniformity observation records of the brightness enhancement composite film; The state calculation module is used to read the reference tension and reference brightness uniformity from the process formula, calculate the tension coupling coefficient based on the reference tension and reference brightness uniformity, and calculate the material segment state of the brightness enhancement composite film by combining the material segment processing record and the tension coupling coefficient. The complete state calculation module is used to arrange the states of the material segments into a one-dimensional sequence, read the states of the material segments from the one-dimensional sequence, calculate the intermediate complete states of the material segments, introduce the local deviation enhancement coefficient, and calculate the final complete state of the material segments in combination with the intermediate complete states. The monitoring and handling module is used to read low and high handling thresholds from a preset database, calculate the handling judgment value corresponding to the current material segment based on the complete status of the material segment, scan sequentially using the segment number of the material segment as an index, construct handling intervals and handling boundaries based on the scanning results, visualize the handling intervals and handling boundaries, and perform intelligent monitoring of the handling intervals.

Images