Intelligent optimization design method and system for greenhouse film formula

Abstract

This invention relates to the field of computer-aided process design technology, specifically to an intelligent optimization design method and system for greenhouse film formulations. The method includes the following steps: collecting signals of crystallinity, amorphous ratio, and crystalline inter-region spacing through film-making experiments and constructing a structure vector; completing crystallization discrimination based on transmittance targets and spacing thresholds; forming a formulation state sequence by combining raw material mass fractions and additive ratios; analyzing the crystallization evolution direction under multiple cooling and traction rate conditions; imposing constraints on parameter value ranges; and outputting the optimized greenhouse film formulation results. In this invention, through the joint characterization and discrimination of multi-dimensional signals of the crystal structure, formulation decisions rely on the microstructural state, combining the crystallization evolution direction and state transition relationship to form dynamic constraints on the raw material mass fraction and additive ratio, ensuring that the optimization results conform to processing conditions, reducing static parameter deviations, maintaining controllable performance and predictable trends under multiple operating conditions, and improving formulation stability and continuous adaptability.

Description

Technical Field

[0001] This invention relates to the field of computer-aided process design technology, and in particular to a method and system for intelligent optimization design of greenhouse film formulations. Background Technology

[0002] Computer-aided process design (CAD) technology primarily involves using computer technology to assist in the design and optimization of product design, process flows, and manufacturing parameters. Core aspects include the automatic generation of process solutions based on computer models, optimization of design parameters, simulation and prediction of process flows, and automated decision-making and scheduling in complex production processes. It covers all stages of product design and plays a particularly important role in industries such as engineering design, manufacturing, chemical engineering, and electronics. By using computer tools and algorithms, design accuracy can be improved, human interference reduced, design solutions optimized, and production efficiency and product quality effectively enhanced.

[0003] The intelligent optimization design method and system for traditional greenhouse film formulations refers to the process of intelligently optimizing and supporting the formulation parameters of greenhouse films using computer technology and intelligent algorithms. Traditional greenhouse film formulation design methods typically rely on manual experience or trial-and-error methods, adjusting the formulation through repeated experiments to meet specific functional requirements. With the advancement of computer technology, computer-aided design-based optimization methods have been gradually introduced. These methods use mathematical models, optimization algorithms, and data-driven approaches to precisely adjust and optimize the material ratios, component characteristics, and process parameters in the formulation, thereby improving design efficiency and reducing experimental costs. Traditional design methods often neglect the comprehensive optimization of multiple objectives, while intelligent optimization design methods can balance multiple objectives, such as cost, strength, and light transmittance, providing the optimal formulation design solution.

[0004] Existing intelligent design methods for greenhouse film formulations rely on empirical rules and static mathematical models to adjust parameters, focusing on the superficial relationship between material ratios and process parameters. This approach fails to reflect the dynamic evolution of the crystalline structure during processing and does not adequately characterize the intrinsic coupling relationship between multi-source detection signals. In scenarios with multiple objective constraints, judgment biases are prone to occur, resulting in high sensitivity of formulation optimization results to actual film-making conditions, insufficient stability and reproducibility. Furthermore, the lack of effective predictive basis when process parameters change can easily lead to performance fluctuations, increase the cost of repeated trials and corrections, and affect design efficiency and product consistency. Summary of the Invention

[0005] To address the technical problems existing in the prior art, embodiments of the present invention provide an intelligent optimization design method for greenhouse film formulations, comprising the following steps:

[0006] S1: Obtain greenhouse film samples through film-making experiments, detect the crystallinity, amorphous ratio, inter-crystalline region signal and transmittance of the greenhouse film samples, calculate the principal component loading coefficient and extract features from all detected signals to generate a crystal structure description vector;

[0007] S2: Based on the crystal structure description vector, the preset transmittance range is converted into a crystallinity range and compared with the crystallinity component. The inter-crystal spacing component is compared with the preset inter-crystal spacing threshold. The comparison results are jointly judged to generate a crystallization discrimination result.

[0008] S3: Obtain the raw material mass fraction and auxiliary agent addition ratio from the preset formula database, filter the raw material mass fraction based on the crystallization discrimination result, and calculate the state transition probability based on the auxiliary agent addition ratio to generate the greenhouse film formula state;

[0009] S4: Under multiple preset cooling rates and traction rates, collect crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state, arrange the crystallinity change sequence and calculate the difference between adjacent terms, analyze the direction of crystallization change, and generate crystallization evolution data.

[0010] S5: Based on the crystallization evolution data, the raw material mass fraction and additive addition ratio constraints corresponding to multiple formulation states are judged, and the value range is limited by the direction of crystallization change to generate the greenhouse film formulation design optimization results.

[0011] As a further aspect of the present invention, the crystal structure description vector includes crystallinity principal component loading, amorphous proportion principal component loading, and inter-crystal spacing signal principal component loading; the crystallization discrimination result includes transmittance interval conformity identifier, inter-crystal spacing threshold conformity identifier, and joint structure discrimination label; the greenhouse film formulation state includes raw material mass fraction combination state index and additive addition ratio state index; the crystallization evolution data includes crystallinity change difference sequence, crystallization change direction identifier sequence, and processing condition state mapping index; and the greenhouse film formulation design optimization result includes raw material mass fraction value range set, additive addition ratio value range set, and multi-formulation state feasible combination set.

[0012] As a further aspect of the present invention, the specific steps of S1 are as follows:

[0013] S101: Obtain greenhouse film samples and perform structural detection on the greenhouse film samples. Collect crystallinity values, amorphous ratio values, inter-crystal spacing signals and transmittance values. Standardize all detection data. Perform principal component decomposition operation based on the discrete distribution of multiple parameters in the sample to generate the principal component sequence of crystal structure.

[0014] S102: Based on the principal component sequence of the crystal structure, perform load solving operation on the contribution relationship of the original parameters corresponding to the multiple principal component components, calculate the load values ​​of crystallinity, amorphous ratio, and intergranular spacing signals under multiple principal components, and vectorize and rearrange the load values ​​to generate principal component load coefficient vector.

[0015] S103: Call the principal component loading coefficient vector, combine it with the standardized crystallinity, amorphous ratio, and intergranular spacing signals, and splice and encapsulate the original signal features and principal component features according to a preset dimension to generate a crystal structure description vector.

[0016] As a further aspect of the present invention, the specific steps of S2 are as follows:

[0017] S201: Based on the crystal structure description vector and extracting the crystallinity component values, the preset transmittance range in the greenhouse film formulation design is converted into a crystallinity range. The crystallinity component values ​​are compared with the upper and lower limits of the crystallinity range to generate a transmittance mapping conformity identifier.

[0018] S202: Based on the transmittance mapping conformity identifier, obtain the preset crystal spacing threshold in the greenhouse film formulation design, call the crystal spacing component and mark the state within or outside the crystal spacing threshold constraint, and serialize and encode the judgment result to generate a crystal spacing threshold judgment identifier.

[0019] S203: Call the transmittance mapping conformity identifier and the crystal region spacing threshold determination identifier, perform joint logic judgment operation on the two types of identifiers, if both types of identifiers are in a conformity state, output the preferred label, otherwise output the rejection label, and generate crystallization discrimination result.

[0020] As a further aspect of the present invention, the crystal region spacing threshold is determined by performing statistics on the crystal region spacing sample sequence obtained during the greenhouse film formulation design stage, averaging multiple crystal region spacing values ​​in the sample sequence and combining them with a combination of three times the standard deviation.

[0021] As a further aspect of the present invention, the specific steps of S3 are as follows:

[0022] S301: Based on the crystallization discrimination result, obtain the raw material mass fraction set and the additive addition ratio, perform index mapping on the multiple mass fraction values ​​in the raw material mass fraction set, obtain the discrimination state and remove the mass fraction values ​​that do not meet the discrimination state constraints, and generate the discrimination constraint raw material mass fraction set.

[0023] S302: Based on the crystallization discrimination result, call the discrimination constraint raw material mass fraction and auxiliary agent addition ratio, combine the raw material mass fraction and auxiliary agent addition ratio, map the mass fraction ratio with the auxiliary agent ratio coefficient, analyze the state weight set and normalize it, and generate the formulation state transition probability set.

[0024] S303: Call the formula state transition probability set, perform sequential connection and state recursion on the transition probabilities between multiple state nodes, serialize and arrange the transition paths according to the probability weights, and integrate the arrangement results to generate the greenhouse film formula state.

[0025] As a further aspect of the present invention, the specific steps of S4 are as follows:

[0026] S401: Collect crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state under multiple cooling rate and traction rate conditions, perform condition identifier binding and sample index alignment on crystallinity under multiple processing conditions, and generate greenhouse film crystallinity change dataset.

[0027] S402: Based on the crystallinity change dataset of the greenhouse film, arrange multiple crystallinity values ​​according to the processing order, perform difference calculation on adjacent crystallinity values, filter out the differences that do not reach the preset crystallinity change discrimination threshold, and associate and map them with the corresponding sequence positions to generate a crystallinity change difference sequence.

[0028] S403: Based on the crystallinity change difference sequence, perform positive and negative discrimination on multiple difference signs, map the discrimination results to the multi-state index in the greenhouse film formulation state, classify and integrate the crystallization change directions corresponding to the multi-state, and generate crystallization evolution data.

[0029] As a further aspect of the present invention, the crystallinity change discrimination threshold is obtained by collecting crystallinity change data of greenhouse film samples under multiple cooling rates and traction rates, performing amplitude distribution calculation on the crystallinity change data, extracting the absolute value sequence of adjacent crystallinity differences, and extracting the median of the differences to determine the threshold.

[0030] As a further aspect of the present invention, the specific steps of S5 are as follows:

[0031] S501: Based on the crystallization evolution data, the multiple formulation state indexes are retrieved item by item, the raw material mass fraction set corresponding to the multiple states is bound to the additive addition ratio, the binding data is checked for consistency, and a formulation parameter constraint mapping set is generated.

[0032] S502: Based on the formula parameter constraint mapping set, for the crystallization change direction associated with multiple formula states, call the crystallization change direction symbol, perform interval determination on the raw material mass fraction and auxiliary agent addition ratio values, mark and remove items that exceed the crystallization change direction limit range, and obtain the parameter value interval set.

[0033] S503: Call the parameter value range set, perform range recombination and combination verification on the raw material mass fraction and auxiliary agent addition ratio under multiple formulation states, summarize the items that meet the preset range consistency rules, and generate the greenhouse film formulation design optimization results.

[0034] The intelligent optimization design system for greenhouse film formulation includes:

[0035] The crystallization characterization module obtains greenhouse film samples through film-making experiments, detects the crystallinity, amorphous ratio, inter-crystal spacing signal and transmittance of the greenhouse film samples, calculates the principal component loading coefficient and extracts features from all detected signals, generates a crystal structure description vector and transmits it to the structure discrimination module.

[0036] The structure discrimination module, based on the crystal structure description vector, converts the preset transmittance range into a crystallinity range and compares the crystallinity component, compares the inter-crystal spacing component with the preset inter-crystal spacing threshold, jointly judges the comparison results, generates a crystallization discrimination result, and transmits it to the formulation modeling module.

[0037] The formula modeling module obtains the raw material mass fraction and auxiliary agent addition ratio from the preset formula database, filters the raw material mass fraction based on the crystallization discrimination result, calculates the state transition probability based on the auxiliary agent addition ratio, generates the greenhouse film formula state, and transmits it to the evolution analysis module.

[0038] The evolution analysis module collects crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state under multiple preset cooling rates and traction rates, arranges the crystallinity change sequence and calculates the difference between adjacent terms, analyzes the direction of crystallization change, generates crystallization evolution data and transmits it to the formulation optimization module.

[0039] The formulation optimization module, based on the crystallization evolution data, constrains the raw material mass fraction and additive addition ratio corresponding to multiple formulation states, and limits the value range by combining the direction of crystallization change, thereby generating the optimized result of greenhouse film formulation design.

[0040] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0041] In this invention, by introducing a joint characterization and discrimination mechanism of multidimensional signals of crystal structure, the formulation decision is based on the microstructural state. Combined with the crystal evolution direction and state transition relationship during processing, dynamic constraints are formed on the raw material mass fraction and the proportion of additives, so that the formulation optimization results are consistent with the actual processing conditions, avoiding deviations caused by single targets or static parameters. Under multiple working conditions, the performance can still be controlled and the trend can be predicted, reducing the dependence on experiments and improving the stability and continuity of formulation adjustment. At the same time, it enhances the comprehensive adaptability of the design results to the requirements of transmission performance and structural balance. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a schematic diagram of the steps of the present invention;

[0044] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0045] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0046] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0047] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0048] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0049] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0050] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0051] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0052] Please see Figure 1 This invention provides an intelligent optimization design method for greenhouse film formulations, comprising the following steps:

[0053] S1: Obtain greenhouse film samples through film-making experiments, detect the crystallinity, amorphous ratio, inter-crystalline region signal and transmittance of the greenhouse film samples, calculate the principal component loading coefficient and extract features from all detected signals to generate a crystal structure description vector;

[0054] S2: Based on the crystal structure description vector, the preset transmittance range is transformed into a crystallinity range and compared with the crystallinity component. The inter-crystal region spacing component is compared with the preset inter-crystal region spacing threshold. The comparison results are jointly judged to generate crystallization discrimination results.

[0055] S3: Obtain the raw material mass fraction and auxiliary agent addition ratio from the preset formula database, filter the raw material mass fraction based on the crystallization discrimination result, and calculate the state transition probability based on the auxiliary agent addition ratio to generate the greenhouse film formula state;

[0056] S4: Under multiple preset cooling rates and traction rates, collect crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state, arrange the crystallinity change sequence and calculate the difference between adjacent terms, analyze the direction of crystallization change, and generate crystallization evolution data.

[0057] S5: Based on crystallization evolution data, the raw material mass fraction and additive addition ratio constraints corresponding to multiple formulation states are judged, and the value range is limited by the direction of crystallization change to generate the optimization results of greenhouse film formulation design.

[0058] The crystal structure description vector includes crystallinity principal component loading, amorphous proportion principal component loading, and inter-crystal spacing signal principal component loading. The crystallization discrimination results include transmittance interval conformity identifier, inter-crystal spacing threshold conformity identifier, and joint structure discrimination label. The greenhouse film formulation status includes raw material mass fraction combination status index and additive addition ratio status index. The crystallization evolution data includes crystallinity change difference sequence, crystallization change direction identifier sequence, and processing condition status mapping index. The greenhouse film formulation design optimization results include raw material mass fraction value range set, additive addition ratio value range set, and multi-formulation status feasible combination set.

[0059] Please see Figure 2 The specific steps of S1 are as follows:

[0060] S101: Obtain greenhouse film samples and perform structural detection on the greenhouse film samples. Collect crystallinity values, amorphous ratio values, inter-crystal spacing signals and transmittance values. Standardize all detection data. Perform principal component decomposition operation based on the discrete distribution of multiple parameters in the sample to generate the principal component sequence of crystal structure.

[0061] The process of obtaining greenhouse film samples and conducting structural testing through film-making experiments is as follows: First, extrusion blow molding process parameters were set, specifically, the extruder temperature was set to 190 degrees Celsius, the blow-up ratio to 2.5, and the cooling air ring velocity to 15 meters per second. Under these process parameters, a polyethylene greenhouse film sample with a thickness of 0.12 mm was prepared. Then, 100 test sample blocks, each 10 mm by 10 mm in size, were cut from the greenhouse film sample. Next, the structure of each of these 100 test sample blocks was tested using an X-ray diffractometer, with the scanning angle range set to 5 degrees to 40 degrees and the scanning rate set to 2 degrees per minute. During the testing of each sample block, the diffractometer collected X-ray intensity signals at different diffraction angles, forming a diffraction pattern curve. For each collected diffraction pattern curve, peak fitting processing was used to decompose the spectrum into crystalline peaks and amorphous dispersed peaks. The areas covered by the crystalline peaks and the amorphous diffuse peaks were calculated separately. The ratio of the crystalline peak area to the total diffraction peak area (i.e., the sum of the crystalline peak area and the amorphous diffuse peak area) was used as the crystallinity value; the ratio of the amorphous diffuse peak area to the total diffraction peak area was used as the amorphous proportion value. Simultaneously, based on Bragg's equation, the interplanar spacing was calculated according to the diffraction angle corresponding to the strongest crystalline peak, and this spacing was used as the inter-crystal spacing signal. For example, for sample block 1, the measured crystalline peak area was 3000 count units, and the amorphous diffuse peak area was 7000 count units. Therefore, its crystallinity value was 3000 / 10000 = 30.0%, and its amorphous proportion value was 7000 / 10000 = 70.0%. The strongest crystalline peak appeared at a diffraction angle of 21.4 degrees, and the calculated inter-crystal spacing signal was 0.415 nanometers. The three types of test data were standardized using the Z-score standardization method. First, the mean crystallinity value was calculated to be 32.0% with a standard deviation of 2.5%, the mean amorphous proportion value was 68.0% with a standard deviation of 2.5%, and the mean intergranular spacing signal was 0.418 nm with a standard deviation of 0.003 nm for 100 samples. For sample block 1, the standardized crystallinity was (30.0-32.0) / 2.5=-0.8; the standardized amorphous proportion was (70.0-68.0) / 2.5=0.8; and the standardized intergranular spacing signal was (0.415-0.418) / 0.003=-1.0. The standardized data from the 100 sample blocks were then used to construct a 100x3 matrix. Principal component decomposition is performed based on the discrete distribution of multiple parameters within the sample. The covariance matrix is ​​calculated for the standardized data matrix of 100 rows and 3 columns, and the eigenvalues ​​and eigenvectors of this covariance matrix are then determined. The eigenvalues ​​are sorted from largest to smallest, and the two eigenvalues ​​with a cumulative contribution rate exceeding 95% are selected. In this embodiment, the two selected eigenvalues ​​are 1.95 and 1.02.The standardized data matrix is ​​multiplied by the eigenvectors corresponding to these two eigenvalues ​​to generate two columns of principal component scores. These two score sequences together constitute the principal component sequence of the crystal structure.

[0062] S102: Based on the principal component sequence of the crystal structure, perform load solving operation on the contribution relationship of the original parameters corresponding to the multiple principal component components, calculate the load values ​​of crystallinity, amorphous ratio, and intergranular spacing signals under multiple principal components, and vectorize and rearrange the load values ​​to generate the principal component load coefficient vector.

[0063] The eigenvalues ​​and corresponding eigenvectors of the calculated covariance matrix are retrieved. Each element of the eigenvector represents the projection of the original parameters onto the corresponding principal component direction, i.e., the initial loading. To obtain more interpretable loadings, each element of the eigenvector is multiplied by the square root of the corresponding eigenvalue. For example, the first eigenvalue obtained in S101 is 1.95, its square root is approximately 1.396, and the corresponding eigenvector is (0.58, -0.58, 0.57); the second eigenvalue is 1.02, its square root is approximately 1.010, and the corresponding eigenvector is (0.41, 0.41, -0.82). Subsequently, the loading values ​​of crystallinity, amorphous proportion, and intergranular spacing signals were calculated under multi-principal component analysis. For the first principal component, the loading value of crystallinity was obtained by multiplying the eigenvector element corresponding to crystallinity (0.58) by the square root of the eigenvalue (1.95), which is 1.396, resulting in 0.81; the loading value of amorphous proportion was obtained by subtracting -0.58 * 1.396 = -0.81; and the loading value of the intergranular spacing signal was obtained by subtracting 0.57 * 1.396 = 0.80. For the second principal component, the loading value of crystallinity was obtained by multiplying 0.41 by the square root of the eigenvalue (1.02), which is 1.010, resulting in 0.41; the loading value of amorphous proportion was obtained by subtracting 0.41 * 1.010 = 0.41; and the loading value of the intergranular spacing signal was obtained by subtracting -0.82 * 1.010 = -0.83. Next, the calculated six load values ​​are vectorized and rearranged according to a fixed order: "crystallinity load of the first principal component, amorphous proportion load of the first principal component, intergranular spacing signal load of the first principal component, crystallinity load of the second principal component, amorphous proportion load of the second principal component, and intergranular spacing signal load of the second principal component". The six values ​​0.81, -0.81, 0.80, 0.41, 0.41, and -0.83 are arranged sequentially to generate a six-dimensional principal component load coefficient vector, namely (0.81, -0.81, 0.80, 0.41, 0.41, -0.83).

[0064] S103: Call the principal component loading coefficient vector, combine it with the standardized crystallinity, amorphous ratio, and intergranular spacing signals, and splice and encapsulate the original signal features and principal component features according to the preset dimensions to generate a crystal structure description vector.

[0065] The generated principal component loading coefficient vector is called, which is (0.81, -0.81, 0.80, 0.41, 0.41, -0.83). Based on the loading correspondence, a linear weighted combination operation is performed on the crystallinity, amorphous ratio, and intercrystalline spacing signals of any greenhouse film sample to be tested. Taking a new greenhouse film sample as an example, its raw data are obtained by X-ray diffraction and calculation, as shown in Table 1 below.

[0066] Table 1. Original data of the greenhouse film samples to be tested

[0067]

[0068] Table 1 lists the raw measured values ​​of three key structural parameters of the greenhouse film samples to be tested. First, the same Z-score normalization process as in S101 was performed on these three raw data points, using the mean (crystallinity 32.0%, amorphous proportion 68.0%, intergranular spacing signal 0.418 nm) and standard deviation (crystallinity 2.5%, amorphous proportion 2.5%, intergranular spacing signal 0.003 nm) of the training sample set. The normalized crystallinity was (35.0-32.0) / 2.5=1.2; the normalized amorphous proportion was (65.0-68.0) / 2.5=-1.2; and the normalized intergranular spacing signal was (0.416-0.418) / 0.003=-0.67. Subsequently, a linear weighted combination operation was performed, which consisted of two parts, each corresponding to one of the two principal components. The first part involves summing the element-wise multiplications of the standardized three data points (1.2, -1.2, -0.67) with the first three coefficients of the principal component loading coefficient vector (0.81, -0.81, 0.80), resulting in the formula: 1.2 * 0.81 + (-1.2) * (-0.81) + (-0.67) * 0.80 = 1.41. The second part involves summing the element-wise multiplications of the standardized three data points with the last three coefficients of the principal component loading coefficient vector (0.41, 0.41, -0.83), resulting in the formula: 1.2 * 0.41 + (-1.2) * 0.41 + (-0.67) * (-0.83) = 0.5561. These two combined results are then vectorized along a unified dimension to form a two-dimensional vector (1.41, 0.5561). Finally, the magnitude of this two-dimensional vector is normalized, and its modulus is calculated, which is the square root of the sum of the squares of 1.41 and 0.5561, approximately 1.515. Each component of the vector is then divided by this modulus: 1.41 / 1.515 = 0.93, 0.5561 / 1.515 = 0.37. Ultimately, the crystal structure description vector for this sample is generated as (0.93, 0.37).

[0069] Please see Figure 3 The specific steps of S2 are as follows:

[0070] S201: Based on the crystal structure description vector and extracting the crystallinity component values, the preset transmittance range in the greenhouse film formulation design is converted into a crystallinity range. The crystallinity component values ​​are compared with the upper and lower limits of the crystallinity range to generate a transmittance mapping conformity identifier.

[0071] The process of obtaining the preset transmittance target range in greenhouse film formulation design based on the crystal structure description vector is as follows: First, the generated crystal structure description vector is called, and the value of this vector is (0.93, 0.37). Since the first principal component has the highest loading weight on the crystallinity parameter in the principal component loading analysis, which is 0.81 and shows a significant positive correlation, the first component of this vector, 0.93, is defined as the crystallinity component value. Subsequently, the operation of obtaining the preset transmittance target range in the greenhouse film formulation design is performed. The target range is set based on the statistical data of historical high transmittance samples. Specifically, 500 high-quality greenhouse film samples with transmittance exceeding 92% are selected as the benchmark set. The crystal structure principal component scores corresponding to all samples in the benchmark set are calculated, and the distribution of the first principal component scores is statistically obtained, with a mean of 1.05 and a standard deviation of 0.10. Based on the principle of 3 standard deviations for a normal distribution, the lower limit of the transmittance target range is set as the mean minus 2 standard deviations, i.e., 1.05 - 0.20 = 0.85; the upper limit is set as the mean plus 2 standard deviations, i.e., 1.05 + 0.20 = 1.25. Therefore, the determined transmittance target range is [0.85, 1.25]. Next, the extracted crystallinity component value of 0.93 is compared with the transmittance target range. The judgment logic is: if the crystallinity component value is greater than or equal to the lower limit of the range and less than or equal to the upper limit of the range, it is considered a match. In this embodiment, the value 0.93 is greater than 0.85 and less than 1.25, satisfying the matching condition. Finally, a transmittance mapping conformance identifier is generated based on the judgment result, and the state of satisfying the matching condition is encoded as the value "1", i.e., a transmittance mapping conformance identifier of "1" is generated.

[0072] S202: Based on the transmittance mapping conformity identifier, obtain the preset crystal spacing threshold in the greenhouse film formulation design, call the crystal spacing component and mark the state within or outside the crystal spacing threshold constraint, and serialize and encode the judgment result to generate a crystal spacing threshold judgment identifier.

[0073] The process of obtaining the preset intergranular spacing threshold in greenhouse film formulation design based on transmittance mapping conforms to the identifier is as follows: First, the second component in the generated crystal structure description vector is identified, i.e., the value 0.37. Reviewing the load analysis, the load weight of the second principal component on the intergranular spacing signal is -0.83, which has the largest absolute value and is significantly correlated. Therefore, this second component 0.37 is defined as the intergranular spacing component. Subsequently, the preset intergranular spacing threshold is obtained. The setting of this threshold aims to constrain the aging performance of the greenhouse film. The specific setting process is as follows: 300 aging-resistant samples that maintained a tensile strength decay rate of less than 10% in accelerated aging tests were collected. The scores of the second principal components of these samples were statistically analyzed, and the numerical distribution range was found to be concentrated between [-0.40, 0.40]. Therefore, -0.40 was set as the lower threshold and 0.40 was set as the upper threshold. Next, the intergranular spacing component 0.37 was called and marked as being within or outside the intergranular spacing threshold constraint. The numerical comparison logic is executed: it is determined whether the inter-regional spacing component 0.37 is between the lower threshold of -0.40 and the upper threshold of 0.40. The comparison shows that 0.37 is greater than -0.40 and less than 0.40, falling within the threshold constraint range. Finally, the judgment result is serialized and encoded, specifying that a value within the threshold range is encoded as "1", and a value outside the range is encoded as "0". In this embodiment, since the value 0.37 is within the range, the inter-regional spacing threshold judgment identifier is generated as "1".

[0074] S203: Call the transmittance mapping conformity flag and the crystal region spacing threshold judgment flag, perform joint logic judgment operation on the two types of flags, if both types of flags are in the conformity state, output the preferred flag, otherwise output the rejection flag, and generate the crystallization discrimination result;

[0075] The process of performing a joint logical judgment operation on the transmittance mapping conformance identifier and the inter-crystal spacing threshold judgment identifier is as follows: First, obtain the output transmittance mapping conformance identifier "1" and the output inter-crystal spacing threshold judgment identifier "1". Then, perform a joint logical judgment operation, which adopts a two-dimensional state mapping mechanism to combine the two input identifier values ​​into a state pair (1, 1). Next, map the joint judgment output state according to a preset discrimination rule. The preset discrimination rule is stored in the logical mapping table shown in Table 2 below.

[0076] Table 2 Mapping Table of Joint Discrimination Rules for Crystallization Structure

[0077]

[0078] As shown in Table 2, this table defines the structural states and discrimination levels corresponding to different identifier combinations. Based on the current input state pair (1, 1), the corresponding row is retrieved in the table, matching the mapping state name "Structural Optimization Qualified" and the corresponding discrimination level code "Level A". Finally, the mapping states are integrated to generate the crystallization discrimination result. In this embodiment, the final output crystallization discrimination result is "Level A (Structural Optimization Qualified)". The effectiveness of this logical deduction has been verified through actual production. Tracking test data of 1000 rolls of greenhouse film show that the greenhouse film judged as "Level A" in this step has an average light transmittance retention rate of up to 95.4% in actual agricultural scenarios, and a damage rate of only 0.5%. Compared with ordinary batches that have not undergone this structural discrimination screening, its light transmittance retention rate has increased by 8.2 percentage points and the damage rate has decreased by 4.1 percentage points, confirming the accuracy and practicality of this joint discrimination logic in screening high-quality structural greenhouse films.

[0079] Please see Figure 4 The specific steps of S3 are as follows:

[0080] S301: Based on the crystallization discrimination result, obtain the raw material mass fraction set and the additive addition ratio, perform index mapping on the multiple mass fraction values ​​in the raw material mass fraction set, obtain the discrimination state and remove the mass fraction values ​​that do not meet the discrimination state constraints, and generate the discrimination constraint raw material mass fraction set.

[0081] The process of obtaining the set of raw material mass fractions for discrimination based on the crystallization discrimination results is as follows: First, the crystallization discrimination results output in the previous step are called. In the previous embodiment, this result has been confirmed as "Grade A", that is, the structural optimization is qualified. Then, the company's internal polyolefin greenhouse film basic formulation database is accessed. This database stores a variety of candidate raw materials and their corresponding mass fraction configurations. For specific screening operations, three typical candidate formulation components are selected as the initial raw material mass fraction sets, labeled as Formulation Component 1, Formulation Component 2, and Formulation Component 3, respectively. Among them, the configuration of Formulation Component 1 is 75.0% linear low-density polyethylene, 20.0% low-density polyethylene, and 5.0% ethylene-vinyl acetate copolymer; the configuration of Formulation Component 2 is 60.0% linear low-density polyethylene, 30.0% low-density polyethylene, and 10.0% ethylene-vinyl acetate copolymer; and the configuration of Formulation Component 3 is 80.0% linear low-density polyethylene, 15.0% low-density polyethylene, and 5.0% ethylene-vinyl acetate copolymer. Next, the index mapping and discrimination status constraint check of multiple mass fraction numerical items are performed. Based on the "Grade A" discrimination result, the preset "high crystallinity structure maintenance constraint condition" is retrieved. This condition clearly stipulates that in order to ensure that the crystallinity is maintained at the optimized qualified level, the mass fraction of the main resin linear low-density polyethylene shall not be less than 70.0%, and the mass fraction of ethylene-vinyl acetate copolymer, which hinders crystallization, shall not be higher than 8.0%. The parameters of formulation components one, two, and three are compared with this constraint condition. For formulation component one, its linear low-density polyethylene mass fraction of 75.0% is greater than 70.0%, and its ethylene-vinyl acetate copolymer mass fraction of 5.0% is less than 8.0%, which meets the constraint and is judged as a retained item. For formulation component two, its linear low-density polyethylene mass fraction of 60.0% is less than the constraint value of 70.0%, and its ethylene-vinyl acetate copolymer mass fraction of 10.0% is greater than the constraint value of 8.0%, which seriously violates the constraint condition and is judged as a rejected item. For formulation component three, its linear low-density polyethylene mass fraction is greater than 70.0% (80.0%) and its ethylene-vinyl acetate copolymer mass fraction is less than 8.0% (5.0%), meeting the constraints and thus being selected as a retained item. Finally, formulation components one and three that meet the criteria are integrated to generate a set of discriminant constraint raw material mass fractions. The advantage of this screening logic is that, through quantitative structure-formulation constraint mapping, formulation combinations that may lead to abnormal crystalline region arrangement or decreased crystallinity are eliminated at the source, ensuring that subsequent processes only optimize within the range of high-quality formulations.

[0082] S302: Based on the crystallization discrimination result, call the discrimination constraint raw material mass fraction and auxiliary agent addition ratio, combine the raw material mass fraction and auxiliary agent addition ratio, map the mass fraction ratio with the auxiliary agent ratio coefficient, analyze the state weight set and normalize it, and generate the formulation state transition probability set.

[0083] The process of generating a formulation state transition probability set based on the crystallization discrimination results is as follows: First, the generated set of discrimination constraint raw material mass fractions, which includes formulation component one and formulation component three, is called. Simultaneously, the additive addition ratios for anti-aging needs are obtained, setting the general hindered amine light stabilizer addition ratio at 2.0%, the anti-dripping and anti-fogging agent addition ratio at 3.0%, and the total additive ratio at 5.0%. Then, the raw material mass fractions and additive addition ratios are jointly calculated to construct a formulation state evaluation model. This model introduces two key influencing factors: the resin matrix crystallization consistency coefficient and the additive dispersion compatibility coefficient. Based on the "Grade A" crystallization discrimination results, the resin matrix crystallization consistency coefficient is set to 0.90, and the additive dispersion compatibility coefficient is set to 0.85. The comprehensive state weight value of each component is calculated. The calculation logic is as follows: the mass fraction of linear low-density polyethylene is multiplied by the resin matrix crystallization consistency coefficient, and the resulting product is then summed with the product of the total additive addition ratio and the additive dispersion compatibility coefficient. Taking formulation component one as an example, the mass fraction 0.75 * 0.90 = 0.675 and the auxiliary agent ratio 0.05 * 0.85 = 0.0425 are summed to obtain the state weight value of formulation component one, which is 0.7175. Taking formulation component three as an example, the mass fraction 0.80 * 0.90 = 0.720 and the auxiliary agent ratio 0.05 * 0.85 = 0.0425 are summed to obtain the state weight value of formulation component three, which is 0.7625. Next, the state weight set is normalized to generate probabilities. First, the sum of the weight values ​​of all retained components is calculated, i.e., 0.7175 + 0.7625 = 1.4800. Then, the transition probabilities of each component are calculated separately: the transition probability of formulation component one is 0.7175 / 1.4800 = 0.4848; the transition probability of formulation component three is 0.7625 / 1.4800 = 0.5152. Finally, a set of recipe state transition probabilities is generated, which includes a probability of 0.4848 for state node one and a probability of 0.5152 for state node three. The specific recipe states and calculation results are shown in Table 3 below.

[0084] Table 3. Calculation Table of Formulation State Weights and Transition Probabilities

[0085]

[0086] As shown in Table 3, the relative advantages of different formulations under the current structural discrimination level are clarified through the above quantitative calculations. Formulation component three, due to its higher linear low-density polyethylene content, exhibits higher state weight and transition probability in maintaining the A-grade crystalline structure. This indicates that in subsequent production adjustments, it will tend to transition to the state of formulation component three.

[0087] S303: Call the formula state transition probability set, perform sequential connection and state recursion on the transition probabilities between multiple state nodes, serialize and arrange the transition paths according to the probability weights, and integrate the arrangement results to generate the greenhouse film formula state.

[0088] The calculated state transition probability sets of the formulations are extracted, namely, the probability of formulation component one is 0.4848 and the probability of formulation component three is 0.5152. To construct the optimal process path that maintains the stability of the crystalline structure during production, a sequential connection of transition probabilities and state recursion operations between multiple state nodes are performed. This process adopts the Markov chain concept, setting the initial production state as the low-probability formulation component one state (as a transition state), and the target steady state as the high-probability formulation component three state (as the final state). Subsequently, the cumulative stability score of the state sequence is calculated. The calculation logic is as follows: the state transition probability of each node in the sequence is weighted and summed with its position weight in the sequence. The position weight is set to increase linearly with the time step, with a weight of 1.0 in the first step and 1.5 in the second step. For the path "state one to state three", its cumulative stability score is: the probability of state one is 0.4848*1.0 + 0.5152*1.5 = 1.2576. If the reverse path "State 3 to State 1" is adopted, its score is 0.5152*1.0+0.4848*1.5=1.2424. By comparing the scores of the two, 1.2576 is greater than 1.2424, therefore "State 1 to State 3" is determined to be the dominant path. Next, this dominant transfer path is sequentially arranged according to probability weights to determine the specific formula adjustment sequence: during the production start-up phase, formula component one (75% linear low-density polyethylene) is used for melt pressure building. After the system stabilizes, it is gradually transitioned to formula component three (80% linear low-density polyethylene) for continuous production. Finally, the arrangement results are integrated to generate the final greenhouse film formula state code as "Sequence-1-3-Optimized". The serialization results were verified on the pilot line. Experimental data showed that when the production was switched using this sequence, the crystal point defect rate of the finished greenhouse film was reduced from 0.8 when using a single formula directly to 0.3, and the standard deviation of the thickness deviation was reduced from 0.005 mm to 0.002 mm, which verified the significant effect of the probability weight-based state sequence in improving processing stability.

[0089] Please see Figure 5 The specific steps of S4 are as follows:

[0090] S401: Collect crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state under multiple cooling rate and traction rate conditions, perform condition identifier binding and sample index alignment on crystallinity under multiple processing conditions, and generate greenhouse film crystallinity change dataset.

[0091] Crystallinity variation data of greenhouse film samples corresponding to different film formulation states were collected under multiple cooling and traction rate conditions. Specifically, firstly, based on the generated film formulation state "Sequence-1-3-Optimized," a dynamic production process transitioning from formulation component one to formulation component three was initiated. To comprehensively capture the structural evolution pattern of this transition stage, multi-dimensional processing windows were set, specifically selecting five discrete time nodes, and adjusting the cooling rate of the cooling air ring and the traction rate of the traction roller respectively. For example, the cooling air velocity of the first node was set to 15.0 m / s and the traction rate to 20.0 m / min; gradually increasing to 25.0 m / s and 30.0 m / min for the fifth node over time. At each set process node, corresponding greenhouse film samples were collected, and their thermal properties were tested using a differential scanning calorimeter. During the test, the heating rate was set to 10.0 degrees Celsius per minute, the nitrogen flow rate was set to 50.0 ml per minute, and the enthalpy of fusion corresponding to the melting peak was recorded. Subsequently, the crystallinity value is calculated by dividing the measured enthalpy of fusion of the sample by the theoretical enthalpy of fusion of fully crystalline polyethylene, 293.0 joules per gram. The quotient is the mass crystallinity of the sample. For example, for the sample collected at the third process node, the measured enthalpy of fusion is 105.5 joules per gram. 105.5 / 293.0 = 0.360, or 36.0%. Next, the condition identifier binding and sample index alignment operation is performed, defining each set of "time index, cooling rate, traction rate" as a unique process condition identifier, and establishing a one-to-one mapping relationship between the calculated crystallinity value and this identifier. The recorded data covers the entire process from the initial state to the steady state, and the specific collected data is shown in Table 4 below.

[0092] Table 4. Data Collection Table of Crystallinity Changes in Greenhouse Film under Dynamic Process Conditions

[0093]

[0094] As shown in Table 4, this table details the melting enthalpy values ​​collected under different process conditions and the corresponding crystallinity calculation results, showing that the crystallinity gradually increases with the increase of process intensity and the change of formulation. Finally, all the aligned data records above are integrated to generate a dataset of greenhouse film crystallinity changes containing complete time-series information. This dataset provides an accurate data foundation for subsequent analysis of structural evolution rate.

[0095] S402: Based on the crystallinity change dataset of greenhouse film, multiple crystallinity values ​​are sorted according to the processing order, and the difference operation is performed on adjacent crystallinity values. Differences that do not reach the preset crystallinity change discrimination threshold are filtered out and associated with the corresponding sequence positions to generate a crystallinity change difference sequence.

[0096] The generated dataset of greenhouse film crystallinity changes is retrieved, and the crystallinity value sequence is extracted according to the time index nodes T1 to T5: 0.316, 0.330, 0.360, 0.393, 0.396. Then, the difference between adjacent crystallinity values ​​is calculated. The logic is as follows: subtract the crystallinity value of the previous time node from the crystallinity value of the later time node to obtain the increment of the interval. The specific calculation process is as follows: the first interval is T2 value 0.330-0.316=0.014; the second interval is T3 value 0.360-0.330=0.030; the third interval is T4 value 0.393-0.360=0.033; the fourth interval is T5 value 0.396-0.393=0.003. Next, a threshold-based filtering operation is performed. A preset threshold for judging crystallinity changes was set to 0.005. This threshold was based on the measurement repeatability error limit of the differential scanning calorimeter. By performing 50 repeated tests on the same standard sample, the standard deviation was measured to be 0.0015. Three times the standard deviation, i.e., 0.0045, was rounded up to 0.005. This aims to eliminate minor fluctuations caused by equipment noise and retain only structural changes with significant physical meaning. The four differences calculated above were compared with this threshold: 0.014 > 0.005, judged as valid changes; 0.030 > 0.005, judged as valid changes; 0.033 > 0.005, judged as valid changes; 0.003 < 0.005, judged as invalid fluctuations and discarded. Finally, the retained valid difference results were correlated and mapped with the corresponding sequence position intervals. The specific mapping results are as follows: the correlation difference between intervals T1 and T2 is 0.014, the correlation difference between intervals T2 and T3 is 0.030, and the correlation difference between intervals T3 and T4 is 0.033. Arranging this cleaned and mapped data in sequence generates a crystallinity change difference sequence. The advantage of this screening logic is that, through quantitative threshold filtering, it accurately identifies the active period of rapid structural evolution, eliminates the interference of minor fluctuations after production stabilizes, and thus accurately pinpoints the key window of influence of formula switching on the structure.

[0097] S403: Based on the crystallinity change difference sequence, perform positive and negative discrimination on multiple difference signs, map the discrimination results to the multi-state index in the greenhouse film formulation state, classify and integrate the crystallization change direction corresponding to the multi-state, and generate crystallization evolution data.

[0098] The output sequence of crystallinity change differences is obtained, namely the values ​​0.014, 0.030, and 0.033. Then, a sign-based discrimination logic is applied to these differences. The discrimination rule is: if the difference is greater than 0, it is marked as "positive growth," representing an increase in crystal region perfection; if the difference is less than 0, it is marked as "negative decline," representing crystal region destruction; if the difference is equal to 0, it is marked as "maintained." In this embodiment, 0.014, 0.030, and 0.033 are all positive numbers, so they are all marked as "positive growth." Next, the discrimination result is indexed and mapped to the greenhouse film formulation state. Reviewing the formulation state sequence, stages T1 to T2 correspond to the mixing period where formulation component one dominates, and stages T2 to T4 correspond to the replacement period transitioning to formulation component three. The difference magnitude is analyzed in conjunction with the formulation stage: in the T1 to T2 interval, a difference of 0.014 corresponds to "positive growth - low speed"; in the T2 to T3 interval, a difference of 0.030 corresponds to "positive growth - acceleration"; and in the T3 to T4 interval, a difference of 0.033 corresponds to "positive growth - high speed". Based on this mapping relationship, a state classification and integration operation is performed, defining three evolution state types: when the difference is less than 0.015 and is positive, it is classified as "induced nucleation state"; when the difference is between 0.015 and 0.040 and is positive, it is classified as "crystal growth state"; and when the difference is greater than 0.040, it is classified as "burst crystallization state". According to this standard, the state in the T1 to T2 interval is classified as "induced nucleation state", and the states in the T2 to T3 and T3 to T4 intervals are classified as "crystal growth state". Finally, the above classification results were packaged to generate the final crystallization evolution data. This data clearly describes the complete physical process by which the internal structure of the greenhouse film underwent induced nucleation followed by rapid crystal growth as the formulation switched from X to Z. The experimental results demonstrate that by monitoring the crystallization evolution status in real time, structurally sensitive areas during formulation switching can be accurately identified. Compared to traditional methods that rely solely on final product testing, this technology can detect structural growth anomalies in advance during production, reducing the response time for process adjustments by 40% and significantly improving the quality stability of transitional products.

[0099] Please see Figure 6 The specific steps of S5 are as follows:

[0100] S501: Based on crystallization evolution data, multiple formulation state indices are retrieved item by item, and the raw material mass fraction sets corresponding to multiple states are bound to the additive addition ratios. Consistency verification is performed on the bound data to generate a formulation parameter constraint mapping set.

[0101] The generated crystallization evolution data is retrieved, and key state nodes are extracted, namely the induced nucleation state corresponding to time interval T1 to T2 and the crystal growth state corresponding to time interval T2 to T4. Simultaneously, the generated greenhouse film formulation state "Sequence-1-3-Optimized" is retrieved to clarify the dynamic evolution logic of the production process, which gradually transitions from formulation component one to formulation component three. Subsequently, a normalization state binding operation is performed on the raw material mass fraction and additive addition ratio. Based on the preset parameters of steps S301 and S302, the design ratio of the basic resin for formulation component one is 75.0% linear low-density polyethylene, 20.0% low-density polyethylene, and 5.0% ethylene-vinyl acetate copolymer, with the total additive addition ratio set at 5.0%. To obtain the actual absolute mass fraction of each component in the final composite of the finished greenhouse film, a normalization calculation operation is performed: the design ratio of each component of the basic resin is multiplied by the effective proportion coefficient of the basic resin in the total formulation, 0.95, which is obtained by subtracting the additive proportion of 0.05 from 1. Taking formulation component one as an example, the absolute mass fraction of linear low-density polyethylene is 0.75 * 0.95 = 0.7125; the absolute mass fraction of low-density polyethylene is 0.20 * 0.95 = 0.1900; the absolute mass fraction of ethylene-vinyl acetate copolymer is 0.05 * 0.95 = 0.0475; and the absolute mass fraction of additives remains at 0.0500. Similarly, calculations are performed on formulation component three, whose basic resin design ratio is 80.0% linear low-density polyethylene, 15.0% low-density polyethylene, and 5.0% ethylene-vinyl acetate copolymer. After normalization, the absolute mass fractions are: linear low-density polyethylene 0.7600, low-density polyethylene 0.1425, ethylene-vinyl acetate copolymer 0.0475, and additives 0.0500. Next, the physical property consistency is checked on the two sets of absolute mass fractions obtained above to ensure the homogeneity of the formulation system in the melt blend state. This check process uses the deviation comparison method between theoretical mixing density and measured density. First, obtain the standard density values ​​of each component: linear low-density polyethylene 0.920 g / cm³, low-density polyethylene 0.923 g / cm³, ethylene-vinyl acetate copolymer 0.940 g / cm³, and additives 0.980 g / cm³. According to the mixing principle, first calculate the ratio of the absolute mass fraction of each component to its corresponding density value, then sum these ratios and take the reciprocal as the theoretical mixing density. For formulation component one, 0.7125 / 0.920 + 0.1900 / 0.923 + 0.0475 / 0.940 + 0.0500 / 0.980 = 1.0820, and take its reciprocal to calculate a theoretical mixing density of approximately 0.9242. Compare this theoretical value with the measured value of 0.9245 g / cm³ obtained on the production line using an online density meter under the same temperature conditions. The absolute value of the difference is 0.0003.A consistency threshold of 0.0010 was set. Since 0.0003 is less than 0.0010, the formulation component configuration was determined to meet the physical consistency requirements, and there was no obvious stratification or segregation. Finally, the absolute mass fractions of each component that passed the verification were associated with the corresponding time points and evolution states and stored to generate a formulation parameter constraint mapping set.

[0102] S502: Based on the formula parameter constraint mapping set, for the crystallization change direction associated with multiple formula states, call the crystallization change direction symbol, perform interval determination on the raw material mass fraction and auxiliary agent addition ratio values, mark and remove items that exceed the crystallization change direction limit range, and obtain the parameter value interval set.

[0103] The generated formula parameter constraint mapping set is read, focusing on the crystal growth state, specifically the changes in key component parameters within the T2 to T4 interval. Within this interval, as the formula switches from component one to component three, the content of linear low-density polyethylene (LLDPE), the main crystallizing agent, increases linearly from 0.7125 to 0.7600. Subsequently, the determined "positive growth" crystallization direction and corresponding "crystal growth state" identifier are invoked. Based on the principles of polymer crystallization kinetics, during the rapid crystal growth stage, excessively high crystalline resin content can easily lead to uncontrolled spherulite size, resulting in increased film haze. Therefore, a dynamic upper limit constraint needs to be set. An upper limit threshold calculation based on the crystallization rate is performed: a base upper limit value of 0.8500 is set, and the calculated maximum crystallinity change difference of 0.033 is used as a dynamic correction factor. The calculation logic is: subtract the product of the dynamic correction factor 0.033 and the sensitivity coefficient 2.0 from the base upper limit value of 0.8500. That is, 0.8500 - 0.066 = 0.7840. Next, a range determination operation is performed, comparing the target content of formulation component three, 0.7600, with the dynamic upper limit of 0.7840. Since 0.7600 is less than 0.7840, this parameter is determined to be within the safe growth range and is retained, allowing the transition to proceed as planned. Simultaneously, a lower limit constraint determination is performed on the additive addition ratio to prevent the additive from being repelled to the surface and causing blooming defects during vigorous crystallization extrusion. The minimum tolerance threshold for the additive in the crystal growth state is set to 0.0450. The current absolute mass fraction of the additive, 0.0500, is compared with this threshold; 0.0500 is greater than 0.0450, thus passing the determination. If, in a simulation or formulation adjustment, the linear low-density polyethylene content is incorrectly set to 0.7900, because it is greater than 0.7840, this parameter will be automatically marked as an "out-of-bounds risk item" and removed from the feasible region, forcing the program to revert to the safe boundary. Finally, all parameter ranges that pass the determination are summarized to generate a parameter value range set. The experimental results show that by introducing a dynamic range determination mechanism based on crystal growth rate, the problem of haze abrupt change caused by excessively rapid crystallization during the transition stage can be effectively avoided. The measured data shows that the average haze of the greenhouse film with this limited range decreased from 6.5% to 4.8% during the formula switching period, and the optical performance was significantly improved by 26.1%.

[0104] S503: Call the parameter value range set, perform range recombination and combination verification on the raw material mass fraction and auxiliary agent addition ratio under multiple formulation states, summarize the items that meet the preset interval consistency rules, and generate the greenhouse film formulation design optimization results;

[0105] The set of output parameter value intervals is obtained, which confirms the feasibility of varying linear low-density polyethylene (LDPE) within the range of 0.7125 to 0.7600, LDPE within the range of 0.1900 to 0.1425, and keeping the additive constant at 0.0500. Subsequently, interval recombination and combination verification are performed to determine the precise formulation for final production execution. The weighted center method is used to determine the optimal operating point. For LDPE, the weighted average of the values ​​at both ends of the interval is selected as the target setpoint. Considering the target steady state in step S303, i.e., the high performance weight of formulation component three, the weighting factor is set to 0.8, and the initial state weighting factor is 0.2. The calculation logic is: initial value 0.7125*0.2 + 0.7600*0.8 = 0.7505. Similarly, a reverse weighted calculation is performed for LDPE, with the initial value 0.1900*0.2 + 0.1425*0.8 = 0.1520. The ethylene-vinyl acetate copolymer content was kept at 0.0475, and the additive content was kept at 0.0500. Next, the final total quantity consistency rule was checked on this set of recombined optimized parameters, namely 0.7505, 0.1520, 0.0475, and 0.0500. The sum of these four values ​​was 0.7505 + 0.1520 + 0.0475 + 0.0500 = 1.0000, satisfying the total mass fraction conservation rule, proving that the formulation design is mathematically and physically sound. Finally, the determined parameters were output as the final optimized greenhouse film formulation design, and corresponding production feeding instructions were formulated. To verify the practical application effect of this optimization result, the greenhouse film produced based on this optimized formulation was labeled as the "optimized group," and the greenhouse film produced using the simple arithmetic mean, i.e., a linear low-density polyethylene content of 0.7360, was labeled as the "control group." Performance comparison tests were conducted on both, including tensile strength and light transmittance. Specific data are shown in Table 5.

[0106] Table 5. Comparison of greenhouse film performance between optimized and control formulations.

[0107]

[0108] As shown in Table 5, the optimized group, by precisely controlling the content of linear low-density polyethylene to 0.7505 and matching crystal growth constraints, achieved a 3.9 MPa increase in tensile strength, a 2.6% increase in light transmittance, and a significant reduction in haze. This indicates that the optimized formulation design generated by this scheme not only theoretically meets the consistency requirements of structural evolution but also achieves dual optimization of mechanical strength and optical performance in practical applications.

[0109] Please see Figure 7 The intelligent optimization design system for greenhouse film formulation includes:

[0110] The crystallization characterization module obtains greenhouse film samples through film-making experiments, detects the crystallinity, amorphous ratio, inter-crystal spacing signal and transmittance of the greenhouse film samples, calculates the principal component loading coefficient and extracts features from all detected signals, generates a crystal structure description vector and transmits it to the structure discrimination module.

[0111] The structure discrimination module, based on the crystal structure description vector, transforms the preset transmittance range into a crystallinity range and compares the crystallinity component. It also compares the inter-crystal spacing component with the preset inter-crystal spacing threshold, jointly judges the comparison results, generates crystallization discrimination results, and transmits them to the formulation modeling module.

[0112] The formula modeling module obtains the raw material mass fraction and auxiliary agent addition ratio from the preset formula database, filters the raw material mass fraction based on the crystallization discrimination results, calculates the state transition probability based on the auxiliary agent addition ratio, generates the greenhouse film formula state, and transmits it to the evolution analysis module.

[0113] The evolution analysis module collects crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state under multiple preset cooling rates and traction rates, arranges the crystallinity change sequence and calculates the difference between adjacent terms, analyzes the direction of crystallization change, generates crystallization evolution data and transmits it to the formulation optimization module.

[0114] The formulation optimization module, based on crystallization evolution data, constrains the raw material mass fraction and additive addition ratio for multiple formulation states, and limits the value range by combining the direction of crystallization change, thereby generating optimized greenhouse film formulation design results.

[0115] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for intelligent optimization design of greenhouse film formulation, characterized in that, Includes the following steps: S1: Obtain greenhouse film samples through film-making experiments, detect the crystallinity, amorphous ratio, inter-crystalline region signal and transmittance of the greenhouse film samples, calculate the principal component loading coefficient and extract features from all detected signals to generate a crystal structure description vector; S2: Based on the crystal structure description vector, the preset transmittance range is converted into a crystallinity range and the crystallinity component is compared. The inter-crystal spacing component is compared with the preset inter-crystal spacing threshold. The comparison results are jointly judged to generate a crystallization discrimination result. S3: Obtain the raw material mass fraction and auxiliary agent addition ratio from the preset formula database, filter the raw material mass fraction based on the crystallization discrimination result, and calculate the state transition probability based on the auxiliary agent addition ratio to generate the greenhouse film formula state; S4: Under multiple preset cooling rates and traction rates, collect crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state, arrange the crystallinity change sequence and calculate the difference between adjacent terms, analyze the direction of crystallization change, and generate crystallization evolution data. S5: Based on the crystallization evolution data, the raw material mass fraction and additive addition ratio constraints corresponding to multiple formulation states are judged, and the value range is limited by the direction of crystallization change to generate the greenhouse film formulation design optimization results.

2. The intelligent optimization design method for greenhouse film formulation according to claim 1, characterized in that, The crystal structure description vector includes crystallinity principal component loading, amorphous proportion principal component loading, and inter-crystal spacing signal principal component loading. The crystallization discrimination result includes transmittance interval conformity identifier, inter-crystal spacing threshold conformity identifier, and joint structure discrimination label. The greenhouse film formulation status includes raw material mass fraction combination status index and additive addition ratio status index. The crystallization evolution data includes crystallinity change difference sequence, crystallization change direction identifier sequence, and processing condition status mapping index. The greenhouse film formulation design optimization result includes raw material mass fraction value range set, additive addition ratio value range set, and multi-formulation status feasible combination set.

3. The intelligent optimization design method for greenhouse film formulation according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Obtain greenhouse film samples and perform structural detection on the greenhouse film samples. Collect crystallinity values, amorphous ratio values, inter-crystal spacing signals and transmittance values. Standardize all detection data. Perform principal component decomposition operation based on the discrete distribution of multiple parameters in the sample to generate the principal component sequence of crystal structure. S102: Based on the principal component sequence of the crystal structure, perform load solving operation on the contribution relationship of the original parameters corresponding to the multiple principal component components, calculate the load values ​​of crystallinity, amorphous ratio, and intergranular spacing signals under multiple principal components, and vectorize and rearrange the load values ​​to generate principal component load coefficient vector. S103: Call the principal component loading coefficient vector, combine it with the standardized crystallinity, amorphous ratio, and intergranular spacing signals, and splice and encapsulate the original signal features and principal component features according to a preset dimension to generate a crystal structure description vector.

4. The intelligent optimization design method for greenhouse film formulation according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the crystal structure description vector and extracting the crystallinity component values, the preset transmittance range in the greenhouse film formulation design is converted into a crystallinity range. The crystallinity component values ​​are compared with the upper and lower limits of the crystallinity range to generate a transmittance mapping conformity identifier. S202: Based on the transmittance mapping conformity identifier, obtain the preset crystal spacing threshold in the greenhouse film formulation design, call the crystal spacing component and mark the state within or outside the crystal spacing threshold constraint, and serialize and encode the judgment result to generate a crystal spacing threshold judgment identifier. S203: Call the transmittance mapping conformity identifier and the crystal region spacing threshold determination identifier, perform joint logic judgment operation on the two types of identifiers, if both types of identifiers are in a conformity state, output the preferred label, otherwise output the rejection label, and generate crystallization discrimination result.

5. The intelligent optimization design method for greenhouse film formulation according to claim 4, characterized in that, The crystal region spacing threshold is determined by performing statistics on the crystal region spacing sample sequence obtained during the greenhouse film formulation design stage, calculating the mean of multiple crystal region spacing values ​​in the sample sequence and combining it with a combination of three times the standard deviation.

6. The intelligent optimization design method for greenhouse film formulation according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Based on the crystallization discrimination result, obtain the raw material mass fraction set and the additive addition ratio, perform index mapping on the multiple mass fraction values ​​in the raw material mass fraction set, obtain the discrimination state and remove the mass fraction values ​​that do not meet the discrimination state constraints, and generate the discrimination constraint raw material mass fraction set. S302: Based on the crystallization discrimination result, call the discrimination constraint raw material mass fraction and auxiliary agent addition ratio, combine the raw material mass fraction and auxiliary agent addition ratio, map the mass fraction ratio with the auxiliary agent ratio coefficient, analyze the state weight set and normalize it, and generate the formulation state transition probability set. S303: Call the formula state transition probability set, perform sequential connection and state recursion on the transition probabilities between multiple state nodes, serialize and arrange the transition paths according to the probability weights, and integrate the arrangement results to generate the greenhouse film formula state.

7. The intelligent optimization design method for greenhouse film formulation according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Collect crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state under multiple cooling rate and traction rate conditions, perform condition identifier binding and sample index alignment on crystallinity under multiple processing conditions, and generate greenhouse film crystallinity change dataset. S402: Based on the crystallinity change dataset of the greenhouse film, arrange multiple crystallinity values ​​according to the processing order, perform difference calculation on adjacent crystallinity values, filter out the differences that do not reach the preset crystallinity change discrimination threshold, and associate and map them with the corresponding sequence positions to generate a crystallinity change difference sequence. S403: Based on the crystallinity change difference sequence, perform positive and negative discrimination on multiple difference signs, map the discrimination results to the multi-state index in the greenhouse film formulation state, classify and integrate the crystallization change directions corresponding to the multi-state, and generate crystallization evolution data.

8. The intelligent optimization design method for greenhouse film formulation according to claim 7, characterized in that, The crystallinity change discrimination threshold is obtained by collecting crystallinity change data of greenhouse film samples under multiple cooling rates and traction rates. The amplitude distribution of the crystallinity change data is calculated, the absolute value sequence of adjacent crystallinity differences is extracted, and the median of the difference is determined.

9. The intelligent optimization design method for greenhouse film formulation according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Based on the crystallization evolution data, the multiple formulation state indexes are retrieved item by item, the raw material mass fraction set corresponding to the multiple states is bound to the additive addition ratio, the binding data is checked for consistency, and a formulation parameter constraint mapping set is generated. S502: Based on the formula parameter constraint mapping set, for the crystallization change direction associated with multiple formula states, call the crystallization change direction symbol, perform interval determination on the raw material mass fraction and auxiliary agent addition ratio values, mark and remove items that exceed the crystallization change direction limit range, and obtain the parameter value interval set. S503: Call the parameter value range set, perform range recombination and combination verification on the raw material mass fraction and auxiliary agent addition ratio under multiple formulation states, summarize the items that meet the preset range consistency rules, and generate the greenhouse film formulation design optimization results.

10. A smart optimization design system for greenhouse film formulations, characterized in that, The system is used to implement the intelligent optimization design method for greenhouse film formulation according to any one of claims 1-9, and the system includes: The crystallization characterization module obtains greenhouse film samples through film-making experiments, detects the crystallinity, amorphous ratio, inter-crystal spacing signal and transmittance of the greenhouse film samples, calculates the principal component loading coefficient and extracts features from all detected signals, generates a crystal structure description vector and transmits it to the structure discrimination module. The structure discrimination module, based on the crystal structure description vector, converts the preset transmittance range into a crystallinity range and compares the crystallinity component, compares the inter-crystal spacing component with the preset inter-crystal spacing threshold, jointly judges the comparison results, generates a crystallization discrimination result, and transmits it to the formulation modeling module. The formula modeling module obtains the raw material mass fraction and auxiliary agent addition ratio from the preset formula database, filters the raw material mass fraction based on the crystallization discrimination result, calculates the state transition probability based on the auxiliary agent addition ratio, generates the greenhouse film formula state, and transmits it to the evolution analysis module. The evolution analysis module collects crystallinity change data of greenhouse film samples corresponding to the greenhouse film formulation state under multiple preset cooling rates and traction rates, arranges the crystallinity change sequence and calculates the difference between adjacent terms, analyzes the direction of crystallization change, generates crystallization evolution data and transmits it to the formulation optimization module. The formulation optimization module, based on the crystallization evolution data, constrains the raw material mass fraction and additive addition ratio corresponding to multiple formulation states, and limits the value range by combining the direction of crystallization change, thereby generating the optimized result of greenhouse film formulation design.

Images