Intelligent optimization method for slitting parameters of stay wire packaging film

Abstract

The invention relates to the technical field of packaging film slitting, in particular to an intelligent optimization method for slitting parameters of a stay wire packaging film. The method comprises the steps that multiple parameter data in the slitting process and thickness data of a target measuring point of a finished film roll on a rolling shaft are collected in real time; the parameter data comprises a plurality of dimension parameters; acquiring a parameter matrix aligned with the thickness data time; inputting the parameter matrix and the thickness data into a constructed prediction model, and outputting predicted thickness data in the next time period; and the moment when the predicted thickness value appearing for the first time in the predicted thickness data is larger than or equal to the set thickness serves as the slitting time of the packaging film, and the packaging film is slit based on the slitting time. According to the scheme, the thickness of the finished film roll can be predicted timely and accurately, the slitting time can be determined timely and accurately, and the problem of waste during slitting of the stay wire packaging film is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of packaging film slitting, and more particularly to an intelligent optimization method for slitting parameters of a drawstring packaging film. Background Art

[0002] Stretch film (also known as stretch film or wrap) is a plastic film widely used in the packaging industry. It's primarily used for pallet packaging, bundling, and securing goods to protect them from damage during transportation and storage. Its primary material is linear low-density polyethylene (LLDPE), which exhibits high tensile strength, good transparency, and excellent adhesion.

[0003] Wire-drawn packaging film slitting is a key step in the production process, which involves cutting wide film rolls into finished film rolls of specific width and length according to customer requirements to meet the needs of different packaging scenarios.

[0004] When cutting into a specific length of finished film roll, it is generally converted into the length of the finished film roll by monitoring the thickness of the finished film roll.

[0005] In existing technology, the technology for monitoring the thickness of finished film rolls is usually to install ultrasonic sensors on the slitting equipment of the production line to measure the thickness of the finished film roll at all times. When the thickness of the finished film roll reaches the set thickness, the slitting equipment sends the corresponding slitting signal to the control module to slit the packaging film. However, this traditional method has several significant technical drawbacks:

[0006] First, the accuracy of thickness monitoring is affected by many factors, including fluctuations in sensor accuracy, uneven thickness of the film material itself, and equipment operational stability. These uncontrollable factors lead to large errors in thickness measurement data, directly affecting the accuracy of the final film roll length.

[0007] Secondly, the issue of response delay in slitting equipment has not been effectively addressed. There's an inevitable mechanical delay between the control system issuing the slitting command and the actual execution of the slitting action. During this time, the slitting equipment continues to operate, resulting in the finished film roll thickness often exceeding the set value. This not only results in product specifications not meeting customer requirements but also leads to significant waste of raw materials, significantly increasing production costs.

[0008] Therefore, when slitting the packaging film, it is particularly important to avoid waste and loss of the pull-line packaging film. Summary of the Invention

[0009] The purpose of the present invention is to propose an intelligent optimization method for the slitting parameters of a drawstring packaging film, so as to solve the problem in the prior art of inaccurate thickness monitoring of the finished film roll, thereby causing waste and loss of the drawstring packaging film; to this end, the present invention provides a solution in one of the following aspects.

[0010] The present invention provides a method for intelligently optimizing slitting parameters of a drawstring packaging film, comprising:

[0011] Real-time collection of multiple parameter data during the slitting process and thickness data of target measurement points of the finished film roll on the reel; the parameter data includes multiple dimensional parameters;

[0012] Obtain a parameter matrix time-aligned with the thickness data;

[0013] Input the parameter matrix and thickness data into the constructed prediction model and output the predicted thickness data for the next period;

[0014] The moment when the predicted thickness value first appears in the predicted thickness data to be greater than or equal to the set thickness is used as the slitting time of the packaging film, and the packaging film is slitting based on the slitting time;

[0015] The loss function of the prediction model is a weighted mean square error loss; the weight is inversely correlated with the degree of noise in each dimension of the data set used to train the prediction model.

[0016] The above scheme starts from two aspects. On the one hand, by analyzing the influence of parameters such as ambient temperature and slitting speed on the tensile deformation of the film material, considering the time lag relationship between different parameters and thickness changes, it is proposed to establish a multi-dimensional parameter matrix to eliminate time deviation; on the other hand, in view of the noise interference existing in the historical parameter collection process, a weighted loss function is used to suppress the influence of high noise parameters on the prediction model, thereby improving the accuracy of the model; by combining the above two aspects, accurate prediction results can be obtained, and after obtaining the prediction results, the slitting action can be triggered in advance to compensate for the equipment response delay, ensure that the actual thickness does not exceed the set thickness, and reduce the waste of packaging film.

[0017] Optionally, the parameters of the multiple dimensions include: ambient temperature, ambient humidity, slitting speed, and coil tightness.

[0018] Optionally, the process of obtaining a parameter matrix time-aligned with the thickness data is:

[0019] Standardizing the plurality of parameter data and thickness data to obtain standardized parameter data and standard thickness data;

[0020] Calculate the mutual information entropy between any standard thickness and each standard parameter in any dimension in the standard thickness data;

[0021] Calculate the correlation coefficient between the standard thickness data and the parameter sequence in any dimension;

[0022] The mutual information entropy and the correlation coefficient are fused to obtain a fusion value, and the lag time corresponding to the maximum value of the fusion value is selected as the optimal lag time of the corresponding standard parameter;

[0023] Based on the optimal lag time, the parameter sequences in each dimension are time-shifted and interpolated to generate a parameter matrix aligned with the thickness data.

[0024] In the above scheme, a time-aligned parameter matrix is constructed to capture the dynamic coupling effect between parameters, so that the predicted thickness data can be accurately obtained.

[0025] Optionally, the weight W k for: is the mean of the noise levels of the parameters of all dimensions in the kth parameter data, N is the total number of dimensions, and exp() is the exponential function.

[0026] Optionally, the noise level is a ratio of each residual value in the residual term to a maximum residual value, and the residual term is obtained by decomposing parameters of the same dimension at different times through STL decomposition.

[0027] The above method obtains the noise level of each dimension parameter, which provides data support for the subsequent contribution of the parameters of this dimension.

[0028] Optionally, the training process of the prediction model is:

[0029] Acquire a data set, the data set including a plurality of historical parameter data in a historical segmentation process and a label of each historical parameter data, wherein the label is the thickness at the next moment after the moment of each parameter data; the historical parameter data is data after delay processing;

[0030] The training set is input into the prediction model for training, and the loss function is used to calculate the loss value. The parameters of the network prediction model are adjusted using the gradient descent algorithm until the loss value between the output predicted thickness value and the label is less than the threshold or the number of training times reaches the set number. The training is stopped and a trained prediction model is obtained.

[0031] Optionally, the loss function s1 is:

[0032] Among them, W k is the weight of the kth parameter data, is the predicted thickness of the kth parameter data, y k is the label of the kth parameter data, and K is the total number of parameter data.

[0033] The introduction of weights can effectively reduce the interference of high-noise dimensional parameters on the prediction model and avoid overfitting of the model to noisy data.

[0034] Optionally, the prediction model is an LSTM model.

[0035] Optionally, the ambient temperature and humidity are collected by temperature and humidity sensors installed on the slitting equipment, the slitting speed is collected by an encoder installed on the slitting equipment, and the tightness of the coil is collected by a pressure sensor installed on the winding shaft.

[0036] Optionally, the slitting process is a time period from the start of winding of the finished film roll to the slitting time when the film reaches a set thickness.

[0037] The beneficial effects of the present invention are:

[0038] The solution of the present invention effectively solves the problem of thickness prediction deviation caused by single sensor monitoring. It significantly improves the slitting time prediction accuracy through a multi-dimensional collaborative mechanism in which environmental parameters compensate for material deformation, process parameters correct the stacking rate, and mechanical parameters reflect the compression state. At the same time, by predicting future thickness trends and making decisions in advance, it avoids waste of raw materials caused by delayed slitting and ensures that the finished product length meets customer customization requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flowchart schematically illustrates the steps of a method for intelligently optimizing slitting parameters of a drawstring packaging film in this embodiment. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0041] like Figure 1 As shown, a method for intelligently optimizing slitting parameters of a drawstring packaging film in this embodiment includes the following steps:

[0042] Step S1 : collecting multiple parameter data during the slitting process and thickness data of a target measuring point of the finished film roll on the reel in real time.

[0043] Among them, the slitting process refers to the process stage of cutting wide-width film materials into finished film rolls (the time span from the start of the winding mechanism to the execution of cutting by the slitting tool). Specifically, it can be achieved by using sensors to monitor the number of revolutions of the winding shaft and the change in film thickness, and the continuous film material can be divided into independent film rolls through physical cutting actions.

[0044] In this embodiment, multiple parameter data are collected during the slitting process, where each parameter data includes parameters of multiple dimensions, specifically ambient temperature, ambient humidity, slitting speed, coil tightness, thickness, etc.

[0045] Thickness is the outer diameter data of the finished film roll collected, which can be measured using an ultrasonic sensor or a laser thickness gauge.

[0046] Among them, the ambient temperature refers to the temperature parameter of the space where the slitting equipment is located. Specifically, a temperature and humidity sensor can be used to monitor the air temperature in the production workshop in real time. This parameter reflects the impact of the thermal expansion effect of the polymer material on the thickness of the film layer stacking.

[0047] The temperature and humidity sensor can be implemented as a patch-type digital temperature and humidity sensor, mounted directly near the slitting equipment's transmission mechanism to prevent interference from the overall workshop environment on the local processing area. The SHT35 digital sensor is an option, offering a measurement accuracy of ±0.3°C and ±2% RH, with a sampling frequency of 10Hz.

[0048] Ambient humidity refers to the air humidity level in the slitting area, which can be collected synchronously by temperature and humidity sensors. This parameter is used to characterize the impact of moisture absorption and deformation of the film material on the actual roll thickness.

[0049] The slitting speed refers to the linear speed of the film material passing through the slitting tool. It is obtained by measuring the spindle speed with an encoder installed on the drive shaft (such as a Hall effect encoder or an incremental rotary encoder). This parameter directly determines the number of superimposed layers of film material per unit time.

[0050] The tightness of the coil refers to the degree of compaction of the film roll on the winding shaft. It can be indirectly calculated by measuring the radial pressure of the winding shaft using a pressure sensor (such as a thin film piezoresistive sensor). This parameter reflects the dynamic impact of the winding tension on the actual compression state of the film layer.

[0051] By incorporating ambient temperature (e.g., increased ambient temperature causes film expansion), humidity, slitting speed, and coil tightness (inadequate winding tension causes film loosening), the above-mentioned factors can effectively identify the impact of material deformation, process fluctuations, and equipment status on thickness. Step S2: Obtain a parameter matrix time-aligned with the thickness data.

[0052] Although different dimensional parameters are theoretically collected at the same time, in reality, the data of different dimensions are not synchronized and there is a time delay between them. Therefore, it is necessary to obtain parameters of different dimensions that are highly consistent with the thickness data changes.

[0053] Specifically, the process of obtaining the parameter matrix time-aligned with the thickness data is:

[0054] Step S21 , performing standardization processing on the plurality of parameter data and thickness data to obtain standardized standard parameter data and standard thickness data.

[0055] Among them, standardization refers to the operation of converting parameter data of different dimensions into a unified dimensional range. Specifically, it can be achieved by using the Z-score standardization method. By eliminating the difference in data dimensions, the deviation caused by different parameter units in subsequent analysis can be avoided.

[0056] Step S22, calculating the mutual information entropy between any standard thickness in the standard thickness data and each standard parameter in any dimension; calculating the correlation coefficient between the standard thickness data and the parameter sequence in any dimension.

[0057] Among them, mutual information entropy refers to a measure of the nonlinear correlation between two random variables, which can be calculated through joint probability distribution and marginal probability distribution to capture the complex nonlinear relationship between parameters and thickness.

[0058] Specifically, by calculating the mutual information entropy between any standard thickness in the standard thickness data and each standard parameter in any dimension, the mutual information entropy between any standard thickness and multiple standard parameters in the same dimension can be obtained.

[0059] Since the calculation of mutual information entropy is an existing technology, it will not be described here in detail.

[0060] The correlation coefficient refers to a quantitative indicator of the degree of linear correlation between two variables, which can be specifically calculated using the Pearson correlation coefficient to measure the degree of linear correlation between the parameter sequence and the thickness data.

[0061] Specifically, the Pearson correlation coefficient between the standard thickness data and the parameter sequence of any dimension is calculated.

[0062] In step S23, the mutual information entropy and the correlation coefficient are fused to obtain a fusion value, and the lag time corresponding to the maximum value of the fusion value is selected as the optimal lag time corresponding to the standard parameter.

[0063] The fusion of mutual information entropy and correlation coefficient can be achieved by setting weights. Specifically, the weights of mutual information entropy and correlation coefficient can be 0.5 respectively; they can also be set according to actual conditions.

[0064] The lag time corresponding to the maximum value of the fusion value is the time interval between the moment when the parameter corresponding to the maximum value of the fusion value is located and the moment when any standard thickness is located.

[0065] It should be noted that each dimension corresponds to an optimal lag time, and the optimal lag times of different dimensions may be the same or different.

[0066] The optimal lag time is the time delay required for parameter changes to affect thickness, reflecting the actual time offset of each dimension's effect on thickness. This optimal lag time helps identify the true correspondence between different dimension data and thickness data, improving the accuracy of subsequent data analysis.

[0067] Step S24 : Based on the optimal lag time, time-shift and interpolation are performed on the parameter sequence in each dimension to generate a parameter data matrix aligned with the thickness data.

[0068] The above-mentioned time shift and interpolation refers to the operation of filling missing values after shifting the parameter sequence according to the optimal lag time, which can be implemented specifically by linear interpolation or spline interpolation method.

[0069] Specifically, the parameter sequences of each dimension are shifted based on the determined optimal lag time, and interpolation processing is used for the missing time points to generate a continuously aligned parameter matrix, thereby obtaining a parameter data matrix aligned with thickness time, wherein one parameter data in the parameter data matrix includes parameters of multiple dimensions.

[0070] The above parameter matrix strictly matches the thickness data in the time dimension, eliminating the time deviation of data acquisition.

[0071] In this embodiment, by dynamically selecting the optimal lag time and combining the dual evaluation mechanism of mutual information entropy and correlation coefficient, the actual influence delay of parameters such as coil tightness and slitting speed on thickness change can be captured more accurately, and the time offset between multi-source parameters and thickness data can be accurately eliminated to generate high-quality data aligned in time and space. It effectively solves the misalignment problem between data of each dimension and thickness data caused by inconsistent parameter response lag, avoids the delay or advance of the slitting action corresponding to the predicted thickness caused by the mismatch between parameters and thickness timing, and thus reduces excessive loss of packaging film materials.

[0072] Step S3: input the parameter matrix and thickness data into the constructed prediction model, and output the predicted thickness data for the next time period.

[0073] The prediction model constructed in this embodiment is an LSTM (Long Short-Term Memory) model. Of course, as another embodiment, the monitoring model can also be a support vector machine.

[0074] The training process of the prediction model is as follows:

[0075] First, get the dataset.

[0076] The data set in this embodiment includes a plurality of historical parameter data in the historical cutting process and a label of each historical parameter data, wherein the label is the thickness at the next moment after the moment of each parameter data.

[0077] Specifically, the historical parameter data includes parameter data at multiple moments in a historical time period, wherein the parameter data includes multi-dimensional parameters, and parameters of the same dimension at different moments constitute a parameter sequence.

[0078] The multi-dimensional parameters in this embodiment include ambient temperature, ambient humidity, slitting speed, coil tightness (with a pressure sensor installed on the reel), thickness, etc. The historical time period can be the time period from the start of winding of the finished film roll to the time of slitting when the set thickness is reached, that is, a cycle of the finished film roll.

[0079] The historical parameter data is data processed by using a lag time compensation algorithm. The specific processing method is the same as that in step S2, which is used to eliminate the timing misalignment between sensor acquisition and thickness change.

[0080] It should be noted that the above processing of historical parameter data takes into account possible delays between different data, where delays can cause temporal misalignment of input data, thereby causing model prediction bias, misjudgment of causal relationships, or reduced accuracy.

[0081] Next, construct the loss function.

[0082] The loss function is the weighted mean square error loss, and the specific acquisition process is:

[0083] First, calculate the weight of each historical parameter data.

[0084] During the slitting parameter optimization process, parameter data encompassing different dimensions may be subject to varying degrees of noise. For example, the ambient temperature sensor's measurement value may fluctuate periodically due to equipment heating, while the slitting speed parameter may exhibit random noise due to mechanical vibration. Therefore, it is necessary to assess the overall credibility of each parameter to minimize the contribution of noisy parameters to the loss calculation. Specifically, parameters with higher noise levels are automatically assigned lower weights, reducing their impact on model parameter updates.

[0085] Specifically, the weight W k for: is the mean of the noise levels of the parameters of all dimensions in the kth parameter data, N is the total number of dimensions, and exp() is the exponential function.

[0086] The noise level is the ratio of each residual value in the residual term to the maximum residual value, and the residual term is obtained by decomposing the parameters of the same dimension at different times using the STL decomposition method.

[0087] Among them, the STL decomposition method refers to the process of decomposing time series data into trend terms, seasonal terms and residual terms. Specifically, the local weighted regression method can be used to iteratively smooth the parameter sequence to separate the residual components reflecting random fluctuations.

[0088] STL decomposition effectively strips away the trend component of the parameters, allowing noise calculation to be performed directly on the residual term. This approach is particularly suitable for dynamic noise scenarios caused by changes in equipment operating status during the slitting process. Furthermore, by inputting the noise mean into an exponential function, a smoothly decreasing weight value is generated, achieving a nonlinear mapping between noise level and weight coefficient, thus avoiding model instability caused by sudden changes in weight.

[0089] The mean value of the noise level is obtained by respectively obtaining the noise level of the parameters of each dimension in the parameter data at each moment, so as to evaluate the noise situation of the parameter data at each moment.

[0090] For example, when the ambient temperature sensor experiences short-term abnormal fluctuations, the corresponding noise level index increases and the weight decreases. At this time, the influence of the temperature parameter on the prediction result will be weakened.

[0091] Next, obtain the loss function.

[0092] Specifically, the loss function s1 is:

[0093] Among them, W k is the weight of the kth parameter data, is the predicted thickness of the kth parameter data, y k is the label of the kth parameter data, and K is the total number of parameter data.

[0094] Through the above technical solution, the present application can accurately identify the dynamic noise level of each dimension parameter during the slitting process, providing an objective basis for the weight allocation of the prediction model, thereby reducing the interference of sensor measurement errors and abnormal parameter fluctuations on thickness prediction, avoiding premature or delayed slitting operations due to noise, and improving the accuracy of slitting length control. When training the prediction model, the weight ratio of each parameter dimension in the loss function can be dynamically adjusted based on the objective quantitative results of the noise level, suppressing the interference of high-noise parameters on prediction accuracy.

[0095] Then, the training set is input into the prediction model for training, and the loss function is used to calculate the loss value. The parameters of the network prediction model are adjusted using the gradient descent algorithm until the loss value between the output predicted thickness value and the label is less than the threshold or the number of training times reaches the set number. The training is stopped and a trained prediction model is obtained.

[0096] By introducing delayed historical data as input, the accuracy of the prediction model can be improved.

[0097] After obtaining the trained prediction model, the parameter matrix and thickness data are input into the prediction model, and the predicted thickness data for the next time period is output.

[0098] The solution of the present invention effectively solves the problem of thickness prediction deviation caused by noise interference in the prediction model. By suppressing the negative impact of high noise parameters, the accuracy of the prediction of subsequent slitting time is improved, thereby reducing the film roll length deviation and raw material waste caused by slitting time judgment errors.

[0099] Step S4: The moment when the predicted thickness value first appears in the predicted thickness data to be greater than or equal to the set thickness is used as the slitting time of the packaging film, and the packaging film is slitting based on the slitting time.

[0100] In this embodiment, when the predicted thickness reaches the set thickness for the first time in the prediction sequence, the control module immediately sends an action instruction to the slitting tool.

[0101] In the above embodiment, the precise time at which the thickness will reach the set value is predicted in advance based on the temporal variation of the slitting parameters. In the event of a response delay in the slitting equipment, the predicted result can be used to trigger the slitting action in advance (bringing the slitting action forward to the theoretical time point when material accumulation is complete, compensating for the error caused by the equipment response delay). This avoids film roll length deviation and material waste caused by the actual thickness exceeding the set value.

[0102] The solution of the present invention integrates the time lag relationship between different parameters and thickness changes and the influence of data noise on the prediction model to obtain an accurate prediction model. Based on the accurate prediction model, the slitting time corresponding to the accurate predicted thickness data is obtained to trigger the slitting action in advance, ensuring that the actual thickness does not exceed the set thickness, thereby reducing the waste of packaging film.

[0103] In the description of this specification, “a plurality of” means at least two, for example, two, three or more, etc., unless otherwise clearly defined.

[0104] Although this specification has shown and described several embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and substitutions without departing from the idea and spirit of the present invention.

Claims

1. A method for intelligent optimization of slitting parameters of a drawstring packaging film, characterized in that: include: Real-time collection of multiple parameter data during the slitting process and thickness data of the target measurement point of the finished film roll on the reel; The parameter data includes multiple dimensional parameters; Obtain a parameter matrix time-aligned with the thickness data; Input the parameter matrix and thickness data into the constructed prediction model and output the predicted thickness data for the next period; The moment when the predicted thickness value first appears in the predicted thickness data to be greater than or equal to the set thickness is used as the slitting time of the packaging film, and the packaging film is slitting based on the slitting time; The loss function of the prediction model is a weighted mean square error loss; the weight is inversely correlated with the degree of noise in each dimension of the data set used to train the prediction model.

2. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 1, characterized in that: The parameters of the multiple dimensions include at least: ambient temperature, ambient humidity, slitting speed, and coil tightness.

3. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 2, characterized in that: The process of obtaining a parameter matrix time-aligned with the thickness data is as follows: Standardizing the plurality of parameter data and thickness data to obtain standardized parameter data and standard thickness data; Calculate the mutual information entropy between any standard thickness and each standard parameter in any dimension in the standard thickness data; Calculate the correlation coefficient between the standard thickness data and the parameter sequence in any dimension; The mutual information entropy and the correlation coefficient are fused to obtain a fusion value, and the lag time corresponding to the maximum value of the fusion value is selected as the optimal lag time of the corresponding standard parameter; Based on the optimal lag time, the parameter sequences in each dimension are time-shifted and interpolated to generate a parameter matrix aligned with the thickness data.

4. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 2, characterized in that: The weight W k for: is the mean of the noise levels of the parameters of all dimensions in the kth parameter data, N is the total number of dimensions, and exp() is the exponential function.

5. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 4, characterized in that: The noise level is the ratio of each residual value in the residual term to the maximum residual value, and the residual term is obtained by decomposing parameters at different times in the same dimension through STL decomposition.

6. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 1, characterized in that: The training process of the prediction model is: Acquire a data set, the data set including a plurality of historical parameter data in a historical segmentation process and a label of each historical parameter data, wherein the label is the thickness at the next moment after the moment of each parameter data; the historical parameter data is data after delay processing; The training set is input into the prediction model for training, and the loss function is used to calculate the loss value. The parameters of the network prediction model are adjusted using the gradient descent algorithm until the loss value between the output predicted thickness value and the label is less than the threshold or the number of training times reaches the set number. The training is stopped and a trained prediction model is obtained.

7. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 1 or 6, characterized in that: The loss function s1 is: Among them, W k is the weight of the kth parameter data, is the predicted thickness of the kth parameter data, y k is the label of the kth parameter data, and K is the total number of parameter data.

8. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 1, characterized in that: The prediction model is an LSTM model.

9. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 2, characterized in that: The ambient temperature and humidity are collected by temperature and humidity sensors installed on the slitting equipment, the slitting speed is collected by an encoder installed on the slitting equipment, and the tightness of the coil is collected by a pressure sensor installed on the reel.

10. The intelligent optimization method for slitting parameters of a drawstring packaging film according to claim 1, characterized in that: The slitting process is the time period from the start of winding of the finished film roll to the slitting time when the set thickness is reached.

Images