A film bag residual degradation day prediction method based on characteristic wave band screening and related device

Abstract

The application discloses a film bag residual degradation day prediction method based on feature band screening and a related device, and belongs to the technical field of degradable film bag thermal degradation. The method comprises the following steps: acquiring real-time infrared spectrum data of a to-be-tested degradable film bag; extracting feature band data in the real-time infrared spectrum data and inputting the feature band data into a pre-trained degradation day estimation model to obtain a degraded day estimation result; and obtaining a residual degradation day according to the degraded day estimation result and a pre-acquired thermal degradation critical day. The training method of the degradation day estimation model comprises the following steps: acquiring feature band data which is strongly correlated with a degradation day and multi-scale feature band fusion data which is sensitive to degradation from historical infrared spectrum data of the film bag in a thermal degradation process, and combining the feature band data and the multi-scale feature band fusion data with corresponding thermal degradation days to form a sample set; and inputting the sample set into a pre-constructed degradation day estimation model to obtain a trained degradation day estimation model, so that the thermal degradation day is accurately estimated.

Description

Technical Field

[0001] This invention belongs to the field of thermal degradation technology of biodegradable membrane bags, and particularly relates to a method and related device for predicting the remaining degradation days of membrane bags based on characteristic band screening. Background Technology

[0002] Traditional plastic bags, due to their non-degradable nature, have caused long-term and serious environmental pollution problems. Developing and using biodegradable film bags has become one of the effective ways to solve this problem. The main components of biodegradable film bags are polybutylene adipate terephthalate, polylactic acid, and talc, playing an increasingly important role in reducing plastic pollution. Unlike non-degradable plastic bags, biodegradable film bags are prone to thermal degradation during high-temperature storage or transportation, leading to abrupt failure of mechanical properties. Accurately understanding the degradation behavior and mechanism of biodegradable plastic film bags in complex environments is of great significance for continuously optimizing product manufacturing processes and customizing the development of new environmentally friendly products.

[0003] Currently, the main methods for detecting the degradation behavior of biodegradable plastic film bags are as follows:

[0004] 1. Estimation of Compost Degradation Behavior: The national standard GB / T 19277.1 specifies a method for determining the final aerobic biodegradability of materials under controlled composting conditions, using the measurement of released carbon dioxide to assess the biodegradability of the materials. However, this method requires continuous monitoring for 3-6 months, is costly and inefficient, and mainly focuses on the recycling stage after the membrane bag is used. There are few studies on estimating the degradation behavior of the membrane bag during storage, transportation, and use, making it effective only in the later stages of degradation and unable to provide early warning.

[0005] 2. Atomic weight measurement: This method assesses the degree of degradation by measuring the change in molecular weight of the membrane bag during the degradation process. While it provides molecular-level degradation information, it requires complex sample pretreatment and high-precision instruments, resulting in high costs, and it cannot monitor the degradation process in real time.

[0006] 3. Infrared Spectroscopy Testing: Fourier transform infrared (FTIR) spectra of the membrane bags are acquired, and the degradation process is monitored by analyzing changes in characteristic peaks. However, traditional infrared spectroscopy often uses the full wavelength range (400-4000 cm⁻¹). -1 While data modeling covers all chemical bond vibration information, the high dimensionality of the data leads to high computational complexity and significant noise interference. Furthermore, full-band analysis requires complex preprocessing, which is time-consuming and cannot meet the needs of rapid on-site detection. In addition, some studies assess the degree of degradation by monitoring changes in the intensity of a single characteristic peak; however, biodegradable membrane bags are multi-component systems with complex interactions among components during degradation. Relying solely on a single peak can easily overlook the synergistic effect of multiple peaks, resulting in large estimation errors.

[0007] In practical applications, the physical properties of biodegradable film bags are mainly affected by temperature during storage and transportation, leading to a gradual decline. Therefore, developing a method to quickly and accurately estimate the thermal degradation behavior of biodegradable film bags during storage and transportation is of great significance for ensuring product quality and service life. Summary of the Invention

[0008] The purpose of this invention is to provide a method and related apparatus for predicting the remaining degradation days of membrane bags based on feature band screening. By optimizing the feature band screening, key narrow band data and related wide band multi-scale fusion data are used as training data to train the estimation model, which improves the efficiency and accuracy of the estimated number of degradation days. By combining the estimated number of degradation days with the pre-calculated number of degradable days, the accurate remaining degradation days can be predicted.

[0009] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0010] In a first aspect, the present invention provides a method for predicting the remaining degradation days of membrane bags based on characteristic band screening, comprising:

[0011] Acquire real-time infrared spectral data of the biodegradable membrane bag under test during the thermal degradation process;

[0012] Acquire the characteristic band data from the real-time infrared spectral data;

[0013] The characteristic band data is input into a pre-trained degradation days estimation model to obtain the estimated number of days of degradation.

[0014] Based on the estimated number of days of degradation already completed and the pre-obtained critical number of days for thermal degradation, the remaining number of days for degradation is obtained.

[0015] The training method for the degradation days estimation model includes:

[0016] Obtain historical infrared spectrum data of biodegradable membrane bags during the thermal degradation process;

[0017] Acquire characteristic band data that are strongly correlated with the number of degradation days from each historical infrared spectral data in the historical infrared spectral dataset, as well as multi-scale characteristic band fusion data that are sensitive to degradation.

[0018] A historical infrared spectral sample set was formed by combining characteristic band data that are strongly correlated with the number of degradation days and are sensitive to degradation, along with their corresponding thermal degradation days.

[0019] The historical infrared spectral sample set is input into a pre-built degradation days estimation model to obtain a trained degradation days estimation model.

[0020] Optionally, obtaining the historical infrared spectral dataset of the biodegradable membrane bag during the thermal degradation process includes:

[0021] Place biodegradable film bags with consistent composition and thickness in a constant temperature environment to accelerate the thermal degradation process.

[0022] Daily full-band infrared spectral data were collected during the thermal degradation process, and the corresponding degradation days of the infrared spectral data were marked to obtain a historical infrared spectral dataset.

[0023] Optionally, methods for obtaining characteristic band data strongly correlated with the number of degradation days from historical infrared spectral data include:

[0024] The historical infrared spectral data are processed by first-order differentiation to obtain historical infrared spectral data with enhanced characteristic peaks;

[0025] Characteristic peaks are extracted from historical infrared spectral data with enhanced characteristic peaks;

[0026] The correlation coefficient between the intensity of the characteristic peak and the number of degradation days is calculated, and the first-order differential spectrum of the strongly correlated peak is obtained by screening according to the preset correlation coefficient threshold.

[0027] Optionally, the step of calculating the correlation coefficient between the intensity of the characteristic peak and the number of degradation days, and selecting the first-order differential spectrum of strongly correlated peaks according to a preset correlation coefficient threshold, includes:

[0028] The Pearson correlation coefficient and Spearman correlation coefficient between the intensity of each characteristic peak and the number of degradation days were calculated respectively;

[0029] First-order differential characteristic peaks with both Pearson and Spearman correlation coefficients greater than preset thresholds were selected as strongly correlated characteristic band data.

[0030] Optionally, methods for obtaining degradation-sensitive multi-scale characteristic band fusion data from historical infrared spectral data include:

[0031] The historical infrared spectral data is preprocessed to obtain effective infrared spectral data;

[0032] The effective infrared spectral data is processed by first-order differentiation to obtain effective infrared spectral data with enhanced characteristic peaks;

[0033] The degradation-sensitive characteristic bands are extracted and fused from the effective infrared spectral data before and after characteristic peak enhancement to obtain sensitive characteristic band fused data.

[0034] Optionally, the step of extracting and fusing degradation-sensitive characteristic bands from the effective infrared spectral data before and after characteristic peak enhancement to obtain sensitive characteristic band fused data includes:

[0035] Feature bands are extracted from the effective infrared spectral data;

[0036] Calculate the signal-to-noise ratio of each characteristic band, and obtain the original spectrum of the sensitive characteristic bands by filtering according to the preset signal-to-noise ratio threshold;

[0037] Based on the selected sensitive characteristic bands, the first-order differential spectrum of the sensitive characteristic bands is obtained;

[0038] The original spectrum and the first-order differential spectrum of the sensitive characteristic peak are fused to obtain fused data of the sensitive characteristic band.

[0039] Optionally, the pre-built degradation days estimation model includes a fusion of principal component analysis (PCA) and partial least squares regression (PLS) algorithms.

[0040] Optionally, the step of inputting the historical infrared spectral sample set into a pre-constructed degradation days estimation model to obtain a trained degradation days estimation model includes:

[0041] The historical infrared spectrum sample set is divided into a training set and a test set;

[0042] Principal component analysis (PCA) is used to reduce the dimensionality of the feature band data in the training set, analyze the eigenvalues ​​and corresponding eigenvectors of the feature band data, and extract the principal components.

[0043] The regression relationship between the principal components and their corresponding thermal degradation days was established by partial least squares regression (PLS), and the number of principal components was iteratively optimized by cross-validation to obtain a trained degradation days estimation model.

[0044] The performance of the trained degradation days estimation model was evaluated using a test set.

[0045] Secondly, the present invention provides a device for predicting the remaining degradation days of membrane bags based on characteristic band screening, comprising:

[0046] Real-time infrared spectral data acquisition module: used to acquire real-time infrared spectral data of the biodegradable film bag under test during the thermal degradation process;

[0047] Real-time feature band data extraction module: used to acquire feature band data from the real-time infrared spectral data;

[0048] The module for obtaining the estimated number of days of degradation is used to input the characteristic band data into a pre-trained degradation days estimation model to obtain the estimated number of days of degradation.

[0049] The module for obtaining remaining degradation days is used to obtain the remaining degradation days based on the estimated degradation days and the pre-obtained critical thermal degradation days.

[0050] The training method for the degradation days estimation model includes:

[0051] Obtain historical infrared spectrum data of biodegradable membrane bags during the thermal degradation process;

[0052] Acquire characteristic band data that are strongly correlated with the number of degradation days from each historical infrared spectral data in the historical infrared spectral dataset, as well as multi-scale characteristic band fusion data that are sensitive to degradation.

[0053] A historical infrared spectral sample set was formed by combining characteristic band data that are strongly correlated with the number of degradation days and are sensitive to degradation, along with their corresponding thermal degradation days.

[0054] The historical infrared spectral sample set is input into a pre-built degradation days estimation model to obtain a trained degradation days estimation model.

[0055] Thirdly, the present invention provides a computer storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for predicting the remaining degradation days of membrane bags based on feature band screening as described in any of the first aspects.

[0056] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: By screening key narrow bands strongly correlated with degradation days from historical infrared spectral data, and performing multi-scale extraction and fusion of degradation-sensitive wide-band data, the key narrow band data and the sensitive wide-band multi-scale fused data are used simultaneously as training data to train the degradation days estimation model. This solves the problems of large noise interference and poor stability of single-peak models in full-band modeling. At the same time, it ensures the amount of modeling data and reduces the overfitting problem in the model training process. Using the trained degradation days estimation results, the estimation model can achieve rapid, non-destructive, and real-time evaluation of the thermal degradation behavior of biodegradable membrane bags, thereby achieving rapid and accurate prediction of the remaining degradation days of biodegradable membrane bags. By using bicorrelation analysis and signal-to-noise ratio calculation, key narrow bands and related wide bands are accurately screened from the full band, reducing data dimensionality and suppressing noise. No complex preprocessing steps are required. First-order differential spectroscopy is used to enhance the resolution of characteristic peaks, reducing dependence on manual intervention, reducing the impact of baseline drift, and further improving model accuracy. Attached Figure Description

[0057] Figure 1 The diagram shows a flowchart of the method for predicting the remaining degradation days of membrane bags based on feature band screening in Embodiment 1 of the present invention.

[0058] Figure 2 The flowchart shown is a method for predicting the remaining degradation days of membrane bags based on feature band screening in Embodiment 2 of the present invention.

[0059] Figure 3The image shown is a schematic diagram of the original infrared spectra of biodegradable film bags taken on different sampling dates in one embodiment of the present invention;

[0060] Figure 4 The image shown is a schematic diagram of the first-order differential infrared spectrum of a biodegradable film bag taken on different sampling dates in one embodiment of the present invention;

[0061] Figure 5 The figure shown is a schematic diagram of the fitting results between the estimated value and the actual value of the degradation days estimation model trained based on full-band spectral data in one embodiment of the present invention.

[0062] Figure 6 The diagram shown illustrates the fitting results between the estimated and actual values ​​of a degradation days estimation model trained based on key narrowband data and related wideband fusion data in one embodiment of the present invention. Detailed Implementation

[0063] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0064] Example 1

[0065] like Figure 1 As shown, this embodiment provides a method for predicting the remaining degradation days of membrane bags based on feature band screening, including:

[0066] Acquire real-time infrared spectral data of the biodegradable membrane bag under test during the thermal degradation process;

[0067] Acquire the characteristic band data from the real-time infrared spectral data;

[0068] The characteristic band data is input into a pre-trained degradation days estimation model to obtain the estimated number of days of degradation.

[0069] Based on the estimated number of days of degradation already completed and the pre-obtained critical number of days for thermal degradation, the remaining number of days for degradation is obtained.

[0070] The training method for the degradation days estimation model includes:

[0071] Obtain historical infrared spectrum data of biodegradable membrane bags during the thermal degradation process;

[0072] Acquire characteristic band data that are strongly correlated with the number of degradation days from each historical infrared spectral data in the historical infrared spectral dataset, as well as multi-scale characteristic band fusion data that are sensitive to degradation.

[0073] A historical infrared spectral sample set was formed by combining characteristic band data that are strongly correlated with the number of degradation days and are sensitive to degradation, along with their corresponding thermal degradation days.

[0074] The historical infrared spectral sample set is input into a pre-built degradation days estimation model to obtain a trained degradation days estimation model.

[0075] By selecting key narrow-band data strongly correlated with the number of days of thermal degradation and relevant wide-band multi-scale fusion data sensitive to thermal degradation as training data, the noise interference problem of traditional full-band analysis is solved. While optimizing the selection of feature data, it can effectively reduce the overfitting problem that may occur during model training, greatly improve the estimation accuracy of the estimation model, and thus improve the prediction accuracy of the remaining degradation days.

[0076] Example 2

[0077] Based on Example 1, this example also incorporates the following design.

[0078] like Figure 2 As shown, in this embodiment, the method for predicting the remaining degradation days of the membrane bag based on feature band screening is implemented through the following steps:

[0079] Step 1: Data Acquisition and Preprocessing.

[0080] A batch of composite film bags with a thickness of 0.004 mm and a composition of 70% polybutylene adipate terephthalate (PBAT), 20% talc, and 10% polylactic acid (PLA) were used. Before the experiment, the film bags were cut into strips of 10cm*30cm. To avoid the influence of factors other than temperature, the test materials were placed in sealed bags, which were then placed in a sealed box. Finally, the sealed box was placed in a constant temperature oven at 65 degrees Celsius. One sample was taken out each day for infrared spectroscopy testing, and the testing was carried out continuously for 100 days, resulting in a total of 100 sets of infrared spectral data.

[0081] In this embodiment, a Bruker V80V Fourier transform infrared spectrometer with a spectral range of 400-4000 cm⁻¹ is used. -1 The test was conducted at room temperature. For each sample, the spectrum was collected three times at three test points, and the average value was taken to eliminate random errors. Figure 3 The image shows infrared spectra at different degradation days, including a 669 cm⁻¹ section. -1 Talc peaks. As can be seen from the figure, with the increase of degradation days, the absorption capacity of some characteristic peaks gradually increases, while the absorption capacity of others shows a decreasing trend.

[0082] Step 2: Filter key narrowband data.

[0083] like Figure 4As shown, a first-order derivative of each set of raw infrared spectra was performed using a filter window width of 11 points and a polynomial order of 2 to enhance the resolution of characteristic peaks and suppress baseline drift. Figure 3 The original spectra in the comparison, Figure 4 MIR in the first-order differential spectrum 1457 band (1430-1492 cm) -1 Peak sharpening.

[0084] Sixteen characteristic peaks (426 cm⁻¹) were extracted from each group of first-order differential spectra. -1 669 cm -1 730 cm -1 752 cm -1 873 cm -1 939 cm -1 1209 cm -1 1320 cm -1 1361 cm -1 1409 cm -1 1457 cm -1 1504 cm -1 1577cm -1 1758 cm -1 2873 cm -1 2958 cm -1 The Pearson correlation coefficient and Spearman correlation coefficient were calculated to determine the relationship between intensity and degradation days.

[0085] The Pearson correlation coefficient R is used to measure the linear correlation between two continuous variables, and its value ranges from -1 to 1. In this embodiment, the Pearson correlation coefficient is used to analyze the linear relationship between the intensity of a characteristic peak and the number of days of degradation. For example, characteristic peaks with |R|>0.45 (such as those located at 730 cm⁻¹) are selected. -1 The characteristic peak (R=-0.855) indicates a significant linear correlation between it and the number of degradation days. The calculation formula is as follows:

[0086] ,

[0087] in, and They represent the first The characteristic peak intensity and degradation days of each sample and These represent the average values ​​of characteristic peak intensity and degradation days, respectively. For sample index, This represents the number of samples.

[0088] The Spearman correlation coefficient ρ measures the monotonic correlation (not limited to linear) between two variables. It is calculated based on the ordinal values ​​of the variables and ranges from -1 to 1. In this embodiment, the Spearman correlation coefficient is used to capture non-linear but monotonic relationships. For example, a characteristic peak with |ρ| > 0.35 (located at 1457 cm⁻¹) is selected. -1 The characteristic peak (ρ=0.414) is used to ensure its monotonic correlation with the number of degradation days. The calculation formula is as follows:

[0089] ,

[0090] in, Indicates the first The difference in ranking of samples based on characteristic peak intensity and degradation days. When calculating the Spearman correlation coefficient, the characteristic peak intensity and degradation days must be converted into ranking values ​​(e.g., 1, 2, ..., n).

[0091] This embodiment uses both the Pearson correlation coefficient R and the Spearman correlation coefficient ρ, requiring both to exceed the threshold (e.g., |R|>0.45 and |ρ|>0.35) to balance linearity and monotonicity and improve the robustness of feature band selection.

[0092] Select strong correlation peaks with |R|>0.45 and |ρ|>0.35, and determine the following characteristic bands as sample data for key narrow bands:

[0093] MIR 730 (717-742 cm) -1 (R=-0.855, ρ=-0.742).

[0094] MIR 752 (744-773 cm) -1 (R=0.379, ρ=0.357).

[0095] MIR 1457 (1430-1492 cm) -1 (R=0.459, ρ=0.414).

[0096] MIR 1504 (1492-1513 cm) -1 (R=0.481, ρ=0.391).

[0097] Step 3: Integrate relevant wideband data.

[0098] Each set of raw infrared spectral data was preprocessed, including baseline correction and smoothing, to reduce noise and baseline drift. Then, a first-order differential was applied to the preprocessed raw infrared spectra using an 11-point filter window and a polynomial order of 2. The purpose of this was to: reduce noise and improve the signal-to-noise ratio (SNR), eliminate baseline drift (such as instrument drift or scattering effects), and enhance the resolution of characteristic peaks (facilitating subsequent characteristic band selection).

[0099] Infrared spectroscopy suffers from baseline drift due to sample scattering (e.g., inhomogeneous thin films), instrument background noise, and interference from non-target components. The goal of baseline correction is to fit and remove low-frequency drift in the spectrum while preserving high-frequency characteristic peak information. In this embodiment, polynomial fitting is used for baseline correction, as shown in the following formula:

[0100] ,

[0101] in, It is the fitted baseline. These are polynomial coefficients. It is the index of the degree term of the polynomial. It is the order of the polynomial (in this embodiment, order 5 is chosen).

[0102] The baseline correction process is as follows:

[0103] Select a smooth region (400-500 cm⁻¹) in the optical spectrum without characteristic peaks. -1 Or 3800-4000cm -1 The least squares method was used to fit the polynomial curve, using the original spectrum. Subtract the fitted curve The corrected spectrum is obtained using the following calculation formula:

[0104] ,

[0105] in, This is the corrected spectrum.

[0106] Infrared spectra typically contain high-frequency noise (such as thermal noise and electronic noise). This embodiment selects a high-precision conformal filter Savitzky-Golay smoothing method to reduce random noise and improve the signal-to-noise ratio. The formula is as follows:

[0107] ,

[0108] in, Indicates the location The waveform value after smoothing. It is the normalization factor, usually a positive integer, used to control the degree of smoothing. ,when The larger the value, the more pronounced the smoothing effect. It is half the window width, indicating the center point. Take from both sides In this embodiment, points are used to form a smooth window. =5, total window width It is 11 o'clock. These are weighting coefficients (determined by polynomial fitting) used for each point within the smoothing window. By applying weights and using a polynomial of order 2 (quadratic polynomial fitting), the spectral peak shape is preserved, avoiding excessive smoothing that could lead to peak distortion. This also maintains high computational efficiency, making it suitable for real-time processing. Indicates the location The value of the original waveform, where Indicates that in The offset relative to the center point within the smooth window centered on the center, with a value range of [value missing]. arrive .

[0109] The estimation results of the first-order differential spectrum are better than those of the original spectrum because the first-order differential processing can enhance the resolution of characteristic peaks (separate overlapping peaks), reduce the influence of baseline drift (due to the disappearance of the constant term after differentiation), and reduce background interference (such as the influence of scattered light).

[0110] This embodiment selects the Savitzky-Golay differential method, and the discrete calculation form of the first-order differential is as follows:

[0111] ,

[0112] in, The derivative of the waveform. Indicates the location The value of the waveform, Indicates the location The value of the waveform, It is the sampling interval, that is, the x-coordinate interval (1 cm) between two adjacent sampling points. -1 ).

[0113] This embodiment uses a polynomial to fit the local window data and then calculates the derivative, as shown in the following formula:

[0114] ,

[0115] in, Indicates the location The derivative value of the waveform. For example... Figure 4 As shown, the characteristic peaks in the spectrum are sharper after the first derivative (e.g., 1457 cm⁻¹). -1Baseline drift was suppressed, and the separation of overlapping peaks was more pronounced (e.g., 730 cm⁻¹). -1 and 752 cm -1 ).

[0116] A fusion dataset of relevant feature bands is constructed by combining the original infrared spectrum and its first-order differential spectrum. The principle is as follows: narrow bands (such as specific chemical bond vibration peaks) provide high specificity but may have limited information; wide bands (such as functional group regions) contain more comprehensive molecular environment information but have lower specificity; fusion of the two can achieve information complementarity and improve model robustness. The scale characteristics of different bands reflect changes in molecular structure at different levels; degradation processes involve molecular chain breakage (sensitive to narrow bands) and changes in crystallinity (sensitive to wide bands). Narrow bands generally have a high signal-to-noise ratio but are susceptible to random errors; wide bands are highly noise-resistant but may contain irrelevant information; fusion can optimize the overall signal-to-noise ratio. The selection method for narrow bands is: |R|>0.45 and |ρ|>0.35; the selection method for wide bands is: SNR>10.

[0117] In this embodiment, to address the resolution issue caused by excessive absorption in certain wavelength bands, only three wavelength segments (wave1, 499-981 cm⁻¹) with high resolution and distinct characteristics from the original infrared spectrum and the first-order differential spectrum are selected based on the signal-to-noise ratio. -1 ), wave2 (1299-1700 cm) -1 ), wave3 (1805-3469 cm) -1 The data are then fused together to form sample data for the relevant wideband.

[0118] Step 4: PCA+PLS model construction and parameter optimization.

[0119] The sample data of key narrow bands and related wide bands in each set of infrared spectral data are grouped into arrays, and the number of days of thermal degradation corresponding to each set of characteristic peak intensity data is added to the corresponding array. The arrays are combined to obtain the historical infrared spectral dataset. The 100 sets of data in the historical infrared spectral dataset are divided into training set and test set in a 7:3 ratio.

[0120] This embodiment combines Principal Component Analysis (PCA) and Partial Least Squares Regression (PLS) to extract key features from infrared spectral data and establish a degradation days estimation model. The training set is input into the pre-built PCA+PLS model. The following is a detailed training process, including the input data for PCA and PLS, operational steps, optimization objectives, and solution methods.

[0121] 1. Dimensionality reduction and feature extraction based on PCA:

[0122] (1) Construct the input data matrix The size is n×p, where n is the number of training set samples (e.g., 70 sets of data), and p is the number of feature bands (e.g., key narrow bands + multi-scale fused bands, a total of p variables). All data have been standardized (mean centering + variance normalization).

[0123] (2) Calculate the covariance matrix to obtain the correlation between the characteristic bands. The calculation formula is as follows:

[0124] ,

[0125] in, Let covariance matrix be the variance matrix. Input data matrix The transpose of .

[0126] (3) Calculate the principal component scores (after dimensionality reduction). The calculation formula is as follows:

[0127] ,

[0128] in, Principal component scores after dimensionality reduction from PCA. Indicates the preceding In this embodiment, there are 1 feature vector. .

[0129] 2. Establish a degradation days estimation model based on PLS:

[0130] (1) Calculate the weight vector (maximize the covariance between spectrum and number of days):

[0131] ,

[0132] in, For the weight vector, Principal component score The transpose of the matrix, This refers to the number of days required for degradation.

[0133] (2) Establish the estimation model:

[0134] ,

[0135] in, This is the regression coefficient matrix. The regression coefficient matrix The transpose of the matrix, This is the residual.

[0136] 3. Determine the optimal number of principal components through 10-fold cross-validation.

[0137] In this embodiment, the optimal principal component is determined to be 3 by calculating the minimum estimation error (Root Mean Square Error, RMSE), thus deriving the trained degradation days estimation model, as shown in the following formula:

[0138] ,

[0139] in, This is the actual value of the degradation days. This is an estimated value for the number of days required for degradation.

[0140] like Figure 5 As shown, the slope of the fitting line between the estimated and actual values ​​of the degradation days estimation model trained on full-band infrared spectral data for the test set is 0.956, and the correlation coefficient R between the actual and estimated degradation days is... 2 =0.899, the root mean square error (RMSEP) of the test set is 9.115%. Figure 6 As shown, the slope of the fitting line between the estimated and actual values ​​of the degradation days estimation model trained on the training set data of this embodiment is 0.968, and the correlation coefficient R between the actual degradation days and the estimated degradation days is... 2 =0.968, and the root mean square error (RMSEP) of the test set is 4.163%, which is a significant improvement over the estimation accuracy trained with traditional full-band infrared spectral data.

[0141] Step 5: Model Application.

[0142] Tensile tests were conducted on the biodegradable film bags, and infrared spectra were collected for the samples after each tensile test. The infrared spectral data of the film bags at the mechanical property critical point were obtained from the collected infrared spectral data. The mechanical property critical point spectrum was used as a reference and compared with the infrared spectral data collected in the thermal degradation experiment. Consistent thermal degradation spectra were obtained, thus the thermal degradation days corresponding to the mechanical property critical point were found to be approximately 65 days.

[0143] Infrared spectroscopy was used to acquire real-time infrared spectral data of a batch of biodegradable membrane bags to be estimated. Feature bands were extracted from the acquired infrared spectral data, and each feature band was input into a degradation day estimation model trained based on the training set data of this embodiment to estimate the number of days the batch of biodegradable membrane bags had degraded. Subtracting the number of days of degradation from the number of days of thermal degradation corresponding to the mechanical performance critical point yields the remaining degradation days of the membrane bags to be tested.

[0144] In this embodiment, quantum mechanical calculations are used to chemically assign characteristic peaks, revealing the correlation between characteristic peaks and degradation mechanisms, such as 730 cm⁻¹. -1The in-plane bending vibration of the aromatic ring CH in PBAT leads to a decrease in peak intensity due to the oxidation and breakage of the aromatic epoxide during degradation; 1457 cm⁻¹ -1 The in-plane bending vibration of the methylene (-CH2-) group in PBAT leads to the recombination of the crystalline region after degradation, resulting in an increase in peak intensity. This study verifies the consistency between the model's estimation results and the material's degradation mechanism, establishing a closed-loop verification system of "spectral characteristics - degradation mechanism - estimation results" to enhance the model's interpretability.

[0145] Example 3

[0146] This embodiment provides a device for predicting the remaining degradation days of membrane bags based on characteristic band screening, including:

[0147] Real-time infrared spectral data acquisition module: used to acquire real-time infrared spectral data of the biodegradable film bag under test during the thermal degradation process;

[0148] Real-time feature band data extraction module: used to acquire feature band data from the real-time infrared spectral data;

[0149] The module for obtaining the estimated number of days of degradation is used to input the characteristic band data into a pre-trained degradation days estimation model to obtain the estimated number of days of degradation.

[0150] The module for obtaining remaining degradation days is used to obtain the remaining degradation days based on the estimated degradation days and the pre-obtained critical thermal degradation days.

[0151] The training method for the degradation days estimation model includes:

[0152] Obtain historical infrared spectrum data of biodegradable membrane bags during the thermal degradation process;

[0153] Acquire characteristic band data that are strongly correlated with the number of degradation days from each historical infrared spectral data in the historical infrared spectral dataset, as well as multi-scale characteristic band fusion data that are sensitive to degradation.

[0154] A historical infrared spectral sample set was formed by combining characteristic band data that are strongly correlated with the number of degradation days and are sensitive to degradation, along with their corresponding thermal degradation days.

[0155] The historical infrared spectral sample set is input into a pre-built degradation days estimation model to obtain a trained degradation days estimation model.

[0156] Example 4

[0157] This embodiment provides a computer storage medium storing a computer program. When the computer program is executed by a processor, it implements the method for predicting the remaining degradation days of membrane bags based on feature band screening as described in any step of Embodiment 1 or Embodiment 2.

[0158] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0159] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0160] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0161] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0162] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for predicting the remaining degradation days of membrane bags based on characteristic band screening, characterized in that, include: Acquire real-time infrared spectral data of the biodegradable membrane bag under test during the thermal degradation process; Acquire the characteristic band data from the real-time infrared spectral data; The characteristic band data is input into a pre-trained degradation days estimation model to obtain the estimated number of days of degradation. Based on the estimated number of days of degradation already completed and the pre-obtained critical number of days for thermal degradation, the remaining number of days for degradation is obtained. The training method for the degradation days estimation model includes: Obtain historical infrared spectrum data of biodegradable membrane bags during the thermal degradation process; Acquire characteristic band data that are strongly correlated with the number of degradation days from each historical infrared spectral data in the historical infrared spectral dataset, as well as multi-scale characteristic band fusion data that are sensitive to degradation. A historical infrared spectral sample set was formed by combining characteristic band data that are strongly correlated with the number of degradation days and are sensitive to degradation, along with their corresponding thermal degradation days. The historical infrared spectral sample set is input into the pre-constructed degradation days estimation model to obtain the trained degradation days estimation model; Methods for obtaining degradation-sensitive multi-scale characteristic band fusion data from historical infrared spectral data include: The historical infrared spectral data is preprocessed to obtain effective infrared spectral data; The effective infrared spectral data is processed by first-order differentiation to obtain effective infrared spectral data with enhanced characteristic peaks; The degradation-sensitive characteristic bands are extracted and fused from the effective infrared spectral data before and after characteristic peak enhancement to obtain sensitive characteristic band fused data. The step of extracting and fusing degradation-sensitive characteristic bands from the effective infrared spectral data before and after characteristic peak enhancement to obtain sensitive characteristic band fused data includes: Feature bands are extracted from the effective infrared spectral data; Calculate the signal-to-noise ratio of each characteristic band, and obtain the original spectrum of the sensitive characteristic bands by filtering according to the preset signal-to-noise ratio threshold; Based on the selected sensitive characteristic bands, the first-order differential spectrum of the sensitive characteristic bands is obtained; The original spectrum and the first-order differential spectrum of the sensitive characteristic peak are fused to obtain fused data of the sensitive characteristic band.

2. The method for predicting the remaining degradation days of membrane bags based on characteristic band screening according to claim 1, characterized in that, The acquisition of historical infrared spectral datasets of the biodegradable membrane bag during the thermal degradation process includes: Place biodegradable film bags with consistent composition and thickness in a constant temperature environment to accelerate the thermal degradation process. Daily full-band infrared spectral data were collected during the thermal degradation process, and the corresponding degradation days of the infrared spectral data were marked to obtain a historical infrared spectral dataset.

3. The method for predicting the remaining degradation days of membrane bags based on characteristic band screening according to claim 1, characterized in that, Methods for obtaining characteristic band data that are strongly correlated with the number of degradation days in historical infrared spectral data include: The historical infrared spectral data are processed by first-order differentiation to obtain historical infrared spectral data with enhanced characteristic peaks; Characteristic peaks are extracted from historical infrared spectral data with enhanced characteristic peaks; The Pearson correlation coefficient and Spearman correlation coefficient between the intensity of each characteristic peak and the number of degradation days were calculated respectively; First-order differential characteristic peaks with both Pearson and Spearman correlation coefficients greater than preset thresholds were selected as strongly correlated characteristic band data.

4. The method for predicting the remaining degradation days of membrane bags based on characteristic band screening according to claim 1, characterized in that, The pre-built degradation days estimation model includes a fusion of principal component analysis (PCA) and partial least squares regression (PLS) algorithms.

5. The method for predicting the remaining degradation days of membrane bags based on characteristic band screening according to claim 4, characterized in that, The step of inputting the historical infrared spectral sample set into a pre-constructed degradation days estimation model to obtain a trained degradation days estimation model includes: The historical infrared spectrum sample set is divided into a training set and a test set; Principal component analysis (PCA) is used to reduce the dimensionality of the feature band data in the training set, analyze the eigenvalues ​​and corresponding eigenvectors of the feature band data, and extract the principal components. The regression relationship between the principal components and their corresponding thermal degradation days was established by partial least squares regression (PLS), and the number of principal components was iteratively optimized by cross-validation to obtain a trained degradation days estimation model. The performance of the trained degradation days estimation model was evaluated using a test set.

6. The method for predicting the remaining degradation days of membrane bags based on characteristic band screening according to claim 1, characterized in that, The method for obtaining the critical number of days for thermal degradation includes: Based on the infrared spectral data of the critical point of mechanical properties of the membrane bag obtained in advance, historical infrared spectral data of the membrane bag during thermal degradation process that are consistent with it are obtained by comparison. Based on the historical infrared spectral data obtained through comparison, the corresponding number of days of thermal degradation is obtained.

7. A device for predicting the remaining degradation days of membrane bags based on characteristic band screening, characterized in that, include: Real-time infrared spectral data acquisition module: used to acquire real-time infrared spectral data of the biodegradable film bag under test during the thermal degradation process; Real-time feature band data extraction module: used to acquire feature band data from the real-time infrared spectral data; The module for obtaining the estimated number of days of degradation is used to input the characteristic band data into a pre-trained degradation days estimation model to obtain the estimated number of days of degradation. The module for obtaining remaining degradation days is used to obtain the remaining degradation days based on the estimated degradation days and the pre-obtained critical thermal degradation days. The training method for the degradation days estimation model includes: Obtain historical infrared spectrum data of biodegradable membrane bags during the thermal degradation process; Acquire characteristic band data that are strongly correlated with the number of degradation days from each historical infrared spectral data in the historical infrared spectral dataset, as well as multi-scale characteristic band fusion data that are sensitive to degradation. A historical infrared spectral sample set was formed by combining characteristic band data that are strongly correlated with the number of degradation days and are sensitive to degradation, along with their corresponding thermal degradation days. The historical infrared spectral sample set is input into the pre-constructed degradation days estimation model to obtain the trained degradation days estimation model; Methods for obtaining degradation-sensitive multi-scale characteristic band fusion data from historical infrared spectral data include: The historical infrared spectral data is preprocessed to obtain effective infrared spectral data; The effective infrared spectral data is processed by first-order differentiation to obtain effective infrared spectral data with enhanced characteristic peaks; The degradation-sensitive characteristic bands are extracted and fused from the effective infrared spectral data before and after characteristic peak enhancement to obtain sensitive characteristic band fused data. The step of extracting and fusing degradation-sensitive characteristic bands from the effective infrared spectral data before and after characteristic peak enhancement to obtain sensitive characteristic band fused data includes: Feature bands are extracted from the effective infrared spectral data; Calculate the signal-to-noise ratio of each characteristic band, and obtain the original spectrum of the sensitive characteristic bands by filtering according to the preset signal-to-noise ratio threshold; Based on the selected sensitive characteristic bands, the first-order differential spectrum of the sensitive characteristic bands is obtained; The original spectrum and the first-order differential spectrum of the sensitive characteristic peak are fused to obtain fused data of the sensitive characteristic band.

8. A computer storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for predicting the remaining degradation days of membrane bags based on feature band screening as described in any one of claims 1-6.

Images