A method for evaluation and prediction of soft capsule environmental adaptability

Abstract

The present application relates to the field of information technology, and discloses a method for evaluating and predicting the adaptability of soft capsules to the environment, which comprises the following steps: generating an environmental risk sequence according to the environmental data and the logistics and storage scene data, and obtaining product tolerance analysis results through simulation analysis; predicting the potential damage probability by using the product tolerance analysis results, and determining the storage adaptability of the target product in the proposed sales area; performing an accelerated aging experiment on a product sample according to the composite environmental stress spectrum, collecting aging-related data, and determining stability indicators; constructing a time series data set, analyzing the coupling relationship and lag effect between different indicators in the time series data set; identifying a key acceleration path according to the analysis results, and correcting a shelf life prediction model; and inputting the environmental stress spectrum of different market areas into the corrected shelf life prediction model to calculate the differentiated shelf life prediction value.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for assessing and predicting the environmental adaptability of soft capsules. Background Technology

[0002] In the pharmaceutical and health food industries, product stability evaluation is a core step in ensuring safety and efficacy, directly impacting consumer health and market trust. This is especially true for soft capsules, whose quality changes during storage affect not only the efficacy of the contents but also the protective properties of the outer shell; therefore, accurate prediction of shelf life is crucial. Research in this area has irreplaceable value in accelerating product launches and optimizing formulation design.

[0003] However, current mainstream stability testing methods have revealed significant shortcomings in practical applications. These methods often ignore the complex external conditions that products face in real-world distribution environments, focusing only on certain fixed testing environments and failing to comprehensively reflect the product's real-world performance under different climate, transportation, and storage scenarios. This limitation leads to a large discrepancy between test results and actual usage variations, making it difficult to meet the needs of rapid development and market adaptation.

[0004] A deeper technical challenge lies in the fact that the stability of soft capsules is not only affected by changes in the chemical composition of the contents, but also closely related to the degradation of the physical properties of the outer shell. As the first line of defense protecting the contents, the aging rate of the outer shell under different environments directly affects the preservation effect of the internal components. For example, the outer shell may soften or leak in humid and hot environments, and such physical changes will accelerate the decomposition of the contents, thus affecting the overall quality. This complexity of the interaction between internal and external factors makes it impossible to fully reveal the causes of product failure by focusing on only one aspect, and it is also difficult to accurately judge its durability in real-world environments.

[0005] Therefore, accurately reproducing the various environmental stresses that soft capsules may experience during their shelf life in testing, while simultaneously considering the interaction between the contents and the outer shell, becomes a key issue for improving prediction accuracy. Solving this problem will directly affect the product's applicability and reliability in different market environments. Summary of the Invention

[0006] This invention provides a method for assessing and predicting the environmental adaptability of soft capsules, mainly comprising:

[0007] The process involves acquiring environmental data and logistics / warehousing scenario data for the target product's intended sales area. The environmental data includes climate-related indicators, and the logistics / warehousing scenario data includes vibration-related indicators. An environmental risk sequence is generated based on the environmental and logistics / warehousing scenario data, and product tolerance analysis results are obtained through simulation analysis. The product tolerance analysis results are used to predict the potential damage probability and determine the storage adaptability of the target product in the intended sales area. A composite environmental stress spectrum with time-series characteristics is generated, simulating the environmental pressure combination of the product from production to the consumer end. Accelerated aging experiments are conducted on product samples based on the composite environmental stress spectrum, collecting aging-related data and determining stability indicators. A time-series dataset is constructed, and the coupling relationship and hysteresis effect between different indicators in the time-series dataset are analyzed. Based on the analysis results, key acceleration paths are identified, and the shelf-life prediction model is revised. Using the revised shelf-life prediction model, the environmental stress spectra of different market areas are input to calculate differentiated shelf-life prediction values. Furthermore, the acquisition of environmental data and logistics warehousing scenario data for the target product's intended sales area includes: extracting historical climate data for the intended sales area from a preset meteorological database to obtain temperature and humidity sequences and light intensity sequences. The temperature and humidity sequences record daily average values, and the light intensity sequences measure cumulative irradiation duration using sensors. Typical logistics warehousing scenario data is collected using IoT sensors to determine the frequency and amplitude of mechanical vibrations. The mechanical vibration frequency is recorded using accelerometers to capture peak fluctuations during transportation, and the amplitude is obtained through frequency domain signal conversion. The temperature and humidity sequences and the mechanical vibration frequencies are integrated to generate the environmental risk sequence. If the environmental risk sequence exceeds a preset threshold, warehousing parameters are adjusted. The data integration calculates a composite value using a weighted average, where the weighted average is the sum of each sequence value multiplied by a preset weight. A transportation impact assessment is simulated based on the environmental risk sequence to obtain the product tolerance analysis results. The product tolerance analysis results are generated using repeated random sampling to produce distribution characteristics, with the environmental risk sequence as the input and the distribution characteristics as the output. Furthermore, the generation of the composite environmental stress spectrum with time-series characteristics includes: acquiring multi-dimensional environmental data, extracting temperature, humidity, and vibration indicators from production process pressure and end-consumer experience to form an initial dataset; performing time-series feature analysis on the initial dataset using an environmental simulation algorithm, and determining the data dimension analysis results by extracting sequence patterns and correlation indicators; generating a stress spectrum generation sequence based on the data dimension analysis results, and integrating multi-dimensional simulation elements; constructing a pressure combination generation model for the stress spectrum generation sequence, and outputting the composite environmental stress spectrum with time-series characteristics by integrating sequence elements and pressure indicators, wherein the composite environmental stress spectrum characterizes the environmental pressure combination that the product may experience from production to the end-consumer stage.Furthermore, the accelerated aging experiment on the product sample based on the composite environmental stress spectrum, the collection of aging-related data, and the determination of stability indicators include: acquiring the composite environmental stress spectrum; conducting an accelerated aging experiment on the product sample in a controlled experimental chamber by applying the composite environmental stress spectrum to obtain aging process data; simultaneously collecting the concentration data of key chemical components of the contents and the physical performance parameter data of the shell using the aging process data; comparing the concentration and parameter deviations between samples from the collected data to determine batch sample differences; using a preset threshold to judge the concentration change trend and performance degradation degree for the batch sample differences, if the deviation exceeds the preset threshold, it is marked as a significant change, and an integrated dataset is obtained; fitting an aging curve from the integrated dataset, determining the curve equation through data point processing, and obtaining the stability indicators. Furthermore, the construction of the time-series dataset and the analysis of the coupling relationships and lag effects among different indicators in the time-series dataset include: acquiring the content of specific degradation products of the contents and the hardness and permeability coefficient of the shell through acquisition equipment to form initial time-series data; integrating chemical change indicators and physical degradation indicators using a weighted average fusion method based on the initial time-series data to obtain a standardized dataset; if the standardized dataset shows a correlation between indicators, determining the degradation trend monitoring value through correlation calculation; acquiring the degradation trend monitoring value, identifying outliers, and optimizing through interpolation to obtain a complete time-series dataset; and processing the complete time-series dataset using an association analysis model to analyze the coupling strength between the chemical change rate and the physical degradation rate under different environmental stress conditions, wherein the coupling strength is quantified based on the confidence value of frequent itemsets. Furthermore, the construction of the time series dataset and the analysis of the coupling relationship and lag effect among different indicators in the time series dataset include: for the initial associated itemset, grouping different environmental stress conditions using conditional variables; the grouping is used to classify rate data through environmental stress condition thresholds to obtain the rate indicator extraction results after grouping; the results extract the change slope as an indicator from the grouped data; from the rate indicator extraction results after grouping, the time series dataset is processed using an association analysis model; frequent itemsets are mined through support and confidence to determine the coupling strength between the chemical change rate and the physical degradation rate; based on the coupling strength, the lag effect delay is obtained; the delay difference under different environmental stress conditions is quantified through relational pattern recognition; the recognition calculates the peak position of the cross-correlation function from the intensity-related time series as the delay; from the delay difference, the effect impact prediction is determined; the prediction is obtained by fitting a regression curve after smoothing the delay data to obtain the coupling relationship and lag effect.Furthermore, the identification of key acceleration paths based on the analysis results includes: for environmental stress conditions, obtaining degradation rate coupling data from the time series dataset, calculating support and confidence using an association analysis model to obtain a significant promoting effect index, where the support is the frequency of itemset occurrence and the proportion of total transactions, and the confidence is the conditional probability; based on the significant promoting effect index, determining the intensity of the physical degradation of the shell on the chemical changes of the contents, and if the intensity exceeds a preset threshold, determining the key acceleration judgment result; from the key acceleration judgment result, obtaining the hysteresis delay, quantifying the delay difference through the peak position of the cross-correlation function, and obtaining degradation threshold monitoring parameters; for the degradation threshold monitoring parameters, using conditional variable grouping to divide the product failure path, obtaining extended analysis results of coupling relationship and hysteresis effect, which are used for subsequent model correction. Furthermore, the revised shelf-life prediction model includes: obtaining the key acceleration path and corresponding environmental stress conditions from a preset environmental database, and determining the path-stress correspondence; obtaining shell state data based on the path-stress correspondence, using the shell state data as the independent variable of the content degradation kinetic equation, and revising the initial shelf-life prediction model, wherein the content degradation kinetic equation takes time and stress as input and outputs a degradation curve; introducing packaging integrity assessment through the revised shelf-life prediction model, wherein the packaging integrity assessment obtains an integrity score by scanning the shell surface, adjusting model parameters, and obtaining an estimated degradation rate; optimizing the prediction accuracy based on the estimated degradation rate, and generating shelf-life prediction results suitable for different environmental conditions. Furthermore, the step of inputting environmental stress spectra of different market regions into the modified shelf-life prediction model to calculate differentiated shelf-life prediction values ​​includes: acquiring typical environmental stress spectra of market regions; extracting temperature and humidity distribution data from a preset database to determine the quantitative indicators of the environmental stress spectra; inputting the quantitative indicators into the modified shelf-life prediction model; adjusting model parameters for the effects of temperature and humidity using accelerated life testing methods to obtain region-specific correction coefficients; the accelerated life testing method accelerates product degradation and fits the life distribution by applying stress higher than normal levels; calculating the product degradation rate in each region based on the correction coefficients; the degradation rate is obtained by multiplying the correction coefficients by a standard degradation function; if the degradation rate exceeds a preset threshold, the initial prediction value is reduced to determine the differentiated shelf-life range; outputting the prediction values ​​for each market from the differentiated shelf-life range; generating adjustment quotas by combining supply chain inventory data to obtain the final regional product shelf-life prediction values.Furthermore, the step of using the product tolerance analysis results to predict the potential damage probability and determine the storage suitability of the target product in the intended sales area includes: extracting distribution characteristic data based on the product tolerance analysis results, wherein the distribution characteristic data is generated through simulation analysis to characterize the product's tolerance under different environmental conditions; calculating the potential damage probability based on the distribution characteristic data, wherein the potential damage probability is extracted from the distribution characteristic data through statistical analysis methods; if the potential damage probability exceeds a preset threshold, adjusting the storage condition parameters, wherein the storage condition parameters include temperature control range and humidity control range; and determining the storage suitability of the target product in the intended sales area based on the potential damage probability and the adjusted storage condition parameters, wherein the storage suitability is obtained by weighted calculation of comprehensive environmental risk and tolerance, and is used to guide the formulation of subsequent warehousing and transportation strategies.

[0008] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0009] This invention discloses a method for predicting and optimizing the shelf life of soft capsule products, addressing the multi-dimensional environmental stresses they may face from production to consumption. The core issue is how the coupling effect between the chemical degradation of the contents and the physical degradation of the shell affects shelf life under different market conditions. This invention generates a composite environmental stress spectrum by acquiring historical climate and logistics data for the target region. Accelerated aging tests are conducted on samples in controlled experiments, simultaneously collecting chemical and physical degradation data. A correlation analysis model is used to reveal the coupling relationship and hysteresis effect of the rates of change between the two, identifying key acceleration paths. This allows for the correction of the shelf life prediction model, incorporating the shell state into the degradation kinetic equation, and ultimately outputting differentiated regional shelf life predictions. Through environmental simulation and data-driven analysis, this invention accurately identifies failure paths and optimizes the prediction model, significantly improving the product's quality assurance capabilities under complex environments. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating a method for assessing and predicting the environmental adaptability of soft capsules according to the present invention. Detailed embodiments.

[0011] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0012] like Figure 1 This embodiment of a method for assessing and predicting the environmental adaptability of soft capsules may specifically include:

[0013] S101. Obtain historical climate data and typical logistics and warehousing scenario data of the target product's intended sales area. The data includes temperature, humidity, light intensity, and mechanical vibration frequency and amplitude.

[0014] Historical climate data for the target sales area is obtained from a pre-set meteorological database to obtain temperature and humidity sequences and light intensity sequences. The temperature and humidity sequences are recorded as daily averages using time series data, while the light intensity sequences are measured by cumulative exposure time using an ultraviolet sensor. Typical logistics and warehousing scenario data is collected using IoT sensors to determine the frequency and amplitude of mechanical vibrations. The mechanical vibration frequency is recorded using accelerometers to capture peak fluctuations during transportation, and the amplitude is converted to a frequency domain signal using Fourier transform, which decomposes a time-domain signal into different frequency components. The temperature and humidity sequences and the mechanical vibration frequencies are integrated to obtain an environmental risk sequence. If the environmental risk sequence exceeds a preset threshold, warehousing parameters are adjusted. Data integration involves calculating the composite value of the temperature and humidity sequences and the mechanical vibration frequency using a weighted average, which is achieved by multiplying each sequence value by a preset weight and then summing the results. Based on the environmental risk sequence, a transportation impact assessment is simulated to obtain product tolerance analysis results. These results are generated using Monte Carlo simulation, a method for estimating uncertainty through repeated random sampling. The input is the environmental risk sequence, and the output is the distribution characteristics. The product tolerance analysis results are used to predict the probability of potential damage and determine the storage suitability of the target product in the intended sales area.

[0015] In one implementation, historical climate data of the target product's intended sales region is obtained. First, the specific geographical location of the intended sales region is determined, such as the Southeast Asian market for a particular electronic product. Average temperature and humidity data for the region over the past five years are extracted by accessing meteorological databases, such as publicly available historical records provided by national meteorological agencies.

[0016] Specifically, temperature data includes daily averages and extreme values, while humidity data includes relative and absolute humidity to assess product stability in high-temperature and high-humidity environments. Furthermore, data from typical logistics and warehousing scenarios are incorporated to consider the transportation and storage processes of products from production sites to sales locations.

[0017] For example, in maritime logistics scenarios, data on light intensity inside containers is collected, and the average intensity values ​​of natural and artificial light are recorded by sensors installed in warehousing facilities, typically in the range of 0 to 1000 lux, in order to analyze the impact of light on product packaging materials.

[0018] Preferably, for acquiring data on the frequency and amplitude of mechanical vibrations, in a land transportation scenario, vibration sensors are used to monitor vibration parameters during truck movement. The specific process involves fixing the sensor to simulated product packaging and recording vibrations with a frequency range of 1 to 100 Hz, as well as acceleration values ​​with peak amplitudes ranging from 0.5 to 5 grams. This data is derived from historical logistics records or on-site simulation tests, ensuring coverage of both static vibrations during warehousing and dynamic vibrations during transportation.

[0019] In one possible implementation, the aforementioned data is integrated to form a comprehensive dataset. For example, climate data is correlated with logistics scenario data to generate reports for product design optimization. This approach enables a comprehensive assessment of environmental factors in the target area, supporting durability analysis of products under different storage conditions.

[0020] It should be noted that this data acquisition process can be extended to various logistics sub-scenarios, such as air transport. Humidity data focuses on high-altitude, low-humidity environments, while vibration data records the frequency distribution of flight turbulence, thereby enhancing the applicability of the technical solution.

[0021] S102. An environmental simulation algorithm is used to generate a series of composite environmental stress spectra with time series characteristics based on the acquired multi-dimensional environmental data. The stress spectra simulate the combination of environmental pressures that the product may experience from production to consumption.

[0022] Multi-dimensional environmental data is acquired, and temperature, humidity, and vibration indicators are extracted from production process pressures and end-consumer experiences to form an initial dataset. Environmental simulation algorithms are used to perform time-series feature analysis on the initial dataset, and the data dimensionality analysis results are determined by extracting sequence patterns and correlation indicators. Based on the data dimensionality analysis results, a stress spectrum generation sequence is generated, integrating multi-dimensional simulation elements. For the stress spectrum generation sequence, a pressure combination generation model is constructed, and by integrating sequence elements and pressure indicators, a composite environmental stress spectrum with time-series characteristics is output.

[0023] In one implementation, the environmental simulation algorithm first collects multi-dimensional environmental data from various stages of the product lifecycle. This data includes dimensions such as temperature, humidity, vibration, and pressure.

[0024] For example, these parameters are recorded in real time by sensors as the product moves from the production workshop to the warehouse and then to the consumer terminal.

[0025] It should be noted that multi-dimensional environmental data refers to a variety of indicators covering physical, chemical, and mechanical aspects to ensure the comprehensiveness of the simulation.

[0026] Specifically, the data acquisition module can be deployed on production lines and logistics vehicles. The collected data is stored in timestamp format to form an initial dataset. Furthermore, based on the collected data, the environmental simulation algorithm uses data fusion technology to generate a composite environmental stress spectrum.

[0027] For example, the algorithm first normalizes the multidimensional data and then integrates the data using sequence modeling methods such as autoregressive models.

[0028] For example, when processing temperature and vibration data, the algorithm calculates the correlation between various dimensions and combines them into a time series vector to simulate the combined pressures experienced by the product at different time points. This fusion process takes into account the interactive effects between data, such as the vibration amplification effect at high temperatures, thereby generating a spectrum that reflects the actual environmental pressure.

[0029] In one possible implementation, the generation of composite environmental stress spectra emphasizes time-series characteristics.

[0030] Specifically, the algorithm divides the data into time segments, each corresponding to a stage in the product lifecycle, such as mechanical stress during production and impact stress during transportation. Through iterative calculations, the algorithm constructs a spectrum, where the horizontal axis represents time and the vertical axis represents combinations of stress intensities. This spectrum simulates the stress sequences that the product may experience.

[0031] For example, ensuring the continuity and accuracy of the spectrum depends on factors ranging from stable temperatures during production to fluctuating humidity during transportation. The key to this process lies in iterative optimization of the algorithm, which can handle data noise and improve the accuracy of the spectrum.

[0032] Preferably, to enhance the versatility of the simulation, the algorithm parameters can be adjusted to adapt to different product types within the same product supply chain domain.

[0033] For example, for electronic products, the algorithm prioritizes the impact of humidity on circuitry, while for food products, it emphasizes the time series of temperature fluctuations. In this way, the generated stress spectrum covers multiple scenarios, but is limited to the logistics environment of consumer goods from production to the end consumer.

[0034] Understandably, the core of environmental simulation algorithms lies in the modeling process of stress combinations.

[0035] Specifically, the process first identifies the peak and mean values ​​of the data in each dimension, and then uses a weighted summation method to calculate the composite stress value.

[0036] For example, the weights for temperature data are set to 0.4 based on historical experience, vibration to 0.3, and humidity to 0.3. These weights are summed to form the stress points at each time point, and connecting these points creates a spectrum. This modeling ensures the dynamic nature of the spectrum, simulating the cumulative stresses a product might face in real-world environments, such as fatigue buildup from long-term transportation. The advantage of this method is that it provides an environmental durability reference for product design, improving reliability. Furthermore, after generating the stress spectrum, the results can be output using visualization tools.

[0037] For example, the spectrum is presented as a curve, with key time points such as warehousing and transfer points marked. This output supports subsequent product testing planning.

[0038] In one embodiment, for a specific product such as a smartphone, the data collected by the algorithm during the process from factory to user includes constant temperature data in the production workshop and data on bumps and vibrations during transportation. The time series spectrum generated by the algorithm shows the transition from stable to high stress, simulating potential damage risks.

[0039] It should be noted that the algorithm's flexibility allows for the integration of additional dimensions, such as lighting data, to expand applications without changing the core framework.

[0040] For example, in another implementation, for clothing products, the algorithm focuses on the combination of humidity and pressure to generate a spectrum to assess the likelihood of wrinkles or deformation during transportation. Through the above steps, the generated composite environmental stress spectrum can effectively simulate the environmental stress combinations throughout the product lifecycle, providing an objective analytical basis.

[0041] S103. In a controlled experimental chamber, an accelerated aging experiment is conducted on the soft capsule sample based on the generated composite environmental stress spectrum, and the concentration data of key chemical components of the contents and the physical performance parameters of the shell are collected simultaneously.

[0042] A composite environmental stress spectrum is obtained, and accelerated aging experiments are conducted on the soft capsule samples in a controlled experimental chamber using this stress spectrum to obtain aging process data. Using this aging process data, the concentration data of key chemical components in the contents and the physical performance parameters of the shell are simultaneously collected. The concentration and parameter deviations between samples are compared from the collected data to determine batch-to-batch sample differences. For these batch-to-batch sample differences, a preset threshold is used to determine the concentration change trend and the degree of performance degradation. If the deviation exceeds the preset threshold, it is marked as a significant change, and an integrated dataset is obtained. An aging curve is fitted from the integrated dataset, and the curve equation is determined by processing the data points using the least squares method to determine the stability index of the soft capsule samples.

[0043] In one implementation, the controlled experimental chamber employs a sealed design, enabling precise control of internal environmental parameters, including temperature, humidity, light intensity, and vibration frequency, to simulate a composite environmental stress spectrum. This chamber is typically equipped with sensors and controllers to ensure stable experimental conditions. The composite environmental stress spectrum refers to a comprehensive stress model generated based on actual storage and transportation environment data, such as a time-series spectrum integrating factors like high temperature and humidity, ultraviolet radiation, and mechanical vibration. This spectrum is generated through historical environmental data analysis to accelerate the aging process of the soft capsules.

[0044] Specifically, the process of generating the composite environmental stress spectrum first involves collecting environmental exposure data of the soft capsules in practical applications, such as temperature fluctuations during warehouse storage and vibration intensity during transportation. This data is then converted into a stress spectrum, for example, by setting the temperature to periodically vary from 25°C to 40°C, and the humidity from 60% to 90%, and superimposing ultraviolet light and low-frequency vibration. The generation of this spectrum relies on environmental simulation software to calculate the interactive effects of various stress factors, ensuring that the spectrum covers the extreme conditions the soft capsules may encounter.

[0045] It should be noted that this generation method can adjust parameters according to different soft capsule types. For example, for vitamin supplement soft capsules, light stress is emphasized to simulate sunlight exposure.

[0046] Preferably, the chamber is calibrated before placing the soft capsule samples in the experimental chamber to ensure that all control modules operate synchronously. Samples are typically arranged in an array to facilitate uniform exposure to the stress environment. After the accelerated aging experiment is initiated, stress is applied based on the generated composite environmental stress spectrum. For example, a combination of high temperature and high humidity is applied in the first stage for 48 hours to simulate long-term storage; vibration and light are introduced in the second stage for 24 hours to simulate transportation. This accelerated method can compress months of aging effects into a few days. The acceleration factor is estimated using the Arrhenius equation principle, but in practice, specific numerical calculations are avoided; only the timing of stress application is controlled. During the experiment, the degradation rate of the sample contents, such as active ingredients, is accelerated and revealed by the stress spectrum, and changes in the elasticity of the shell also occur. Furthermore, synchronous data acquisition involves installing a multi-channel sensor system, such as using high-performance liquid chromatography to monitor the concentration of key chemical components in the contents online, such as changes in vitamin C content. Physical performance parameters, such as the hardness, elasticity, and thickness of the shell, are collected using a tensile testing instrument. These sensors are connected to the control system of the experimental chamber to achieve real-time data acquisition, recording once per minute to ensure the synchronization of chemical concentration data and physical parameter data.

[0047] In one possible approach, for soft capsules containing oil-based contents, the focus is on the oxidation level of the fat-soluble components, such as monitoring peroxide values ​​as a key concentration indicator, while simultaneously assessing the shell's rupture strength. This approach expands the applicability of the experiment to cover different content types.

[0048] For example, in experiments, if soft capsules are used for fish oil supplementation, composite environmental stress spectrum can preferentially enhance oxidation-related stresses, such as increasing oxygen concentration and temperature cycling, while simultaneously collecting data on the decrease in omega-3 fatty acid concentration and changes in shell sealing parameters, thereby assessing overall stability.

[0049] Understandably, this synchronous acquisition ensures data correlation. For example, when the concentration of chemical components decreases by more than 10%, the trend of increasing physical parameters such as shell brittleness can be captured and used for subsequent analysis.

[0050] In one embodiment, after the experiment, the effectiveness of accelerated aging is verified by data comparison, such as comparing the correlation between experimental data and natural aging data to confirm the accuracy of the composite stress spectrum simulation. Through the above steps, this method can achieve reliable aging testing of soft capsules in the pharmaceutical field, providing data support for optimizing product formulation and packaging design.

[0051] S104. Establish a time series dataset of chemical change indicators of contents and physical degradation indicators of shell, wherein the chemical change indicators include the content of specific degradation products, and the physical degradation indicators include shell hardness and permeability coefficient.

[0052] The content of specific degradation products of the contents and the hardness and permeability of the shell are acquired by the acquisition device to form initial time series data. Based on the initial time series data, a weighted average fusion method is used to integrate chemical change indicators and physical degradation indicators to obtain a standardized dataset. If the standardized dataset shows correlation between indicators, the degradation trend monitoring value is determined by Pearson correlation calculation. The degradation trend monitoring value is obtained, outliers are identified, and interpolation optimization is performed to obtain the complete time series dataset.

[0053] In one implementation, a time-series dataset of indicators of chemical changes in contents and physical degradation of the shell is established, firstly by acquiring data through periodic sampling.

[0054] Specifically, for indicators of chemical changes in the contents, the content of specific degradation products can be measured using chromatographic analysis methods. For example, high-performance liquid chromatography (HPLC) can be used to separate and quantify the content of target compounds. This method achieves separation and detection based on the differences in the distribution of compounds between the stationary and mobile phases, thus obtaining precise values ​​at each time point. Furthermore, physical degradation indicators include shell hardness and permeability coefficient. Shell hardness can be evaluated using a Vickers hardness tester, which involves applying a specific load to the shell surface and measuring the indentation size to calculate the hardness value. The permeability coefficient is measured using a permeation testing device to measure the rate at which gas or liquid passes through the shell, for example, recording the relationship between permeation and time under constant pressure, thereby quantifying the barrier performance of the material. These indicators must be collected under identical environmental conditions to ensure data consistency.

[0055] For example, in the implementation scenario of materials storage, for food packaging materials, the contents may be organic compounds, and the content of their degradation products, such as aldehydes, changes over time. The process of constructing a time-series dataset includes data collection at an initial time point, followed by repeated measurements of chemical and physical indicators at fixed intervals (e.g., weekly), organizing all data into a sequence in chronological order.

[0056] For example, the dataset can be stored in tabular form, with each row corresponding to a timestamp and columns recording the content of degradation products, shell hardness value, and permeability coefficient value, respectively.

[0057] It should be noted that the establishment of time series also involves data preprocessing steps, such as outlier removal and interpolation filling, to handle noise in the measurement.

[0058] Specifically, if data is missing at a certain point in time, it can be estimated using linear interpolation based on adjacent points to ensure the continuity of the sequence. This method helps to capture degradation trends; for example, when the content of degradation products exceeds a threshold, the shell hardness often decreases accordingly, while the permeability increases, thus revealing the overall degradation pattern of the material.

[0059] In one possible implementation, variations of sampling frequencies can be introduced to enhance the versatility of the dataset.

[0060] For example, in industrial materials monitoring, daily sampling is used for packaging shells in high-humidity environments to capture rapid changes, while monthly sampling is used in low-temperature storage scenarios, also limited to the area of ​​material degradation. This flexibility allows the dataset resolution to be adjusted according to the specific application.

[0061] Preferably, the dataset can be further used for trend analysis, such as quantifying the association between chemical changes and physical degradation by calculating the correlation coefficients between indicators. Specifically, this involves applying Pearson correlation analysis to the time series, calculating the covariance of degradation product content and hardness values ​​divided by their respective standard deviations, thus obtaining correlation strength values ​​ranging from -1 to 1. This analysis helps identify key degradation points; for example, a correlation coefficient close to -0.8 indicates that chemical degradation significantly affects the physical structure.

[0062] Understandably, in another embodiment, for pharmaceutical packaging materials, the chemical changes of the contents focus on the degradation products of the active ingredient, such as oxide content, quantified by measuring the intensity of their absorption peaks using a spectrometer; the physical degradation of the outer shell is assessed by evaluating the hardness of the plastic film and the water vapor permeability coefficient, using similar testing methods to collect data and form a time series. This scenario demonstrates the applicability of the technical solution within the same field. Furthermore, the entire dataset creation process emphasizes modularity; for example, the data acquisition module is responsible for measuring indicators, and the sequence construction module integrates time information. Through these steps, comprehensive tracking of the material degradation process is achieved, supporting subsequent predictive model development without introducing subjective optimization.

[0063] For example, in one specific implementation, for battery casing materials, chemical indicators monitor the accumulation of specific degradation products in the electrolyte, while physical indicators assess casing hardness and ion permeability. The dataset covers changes from the initial state to several months later, revealing degradation patterns under long-term storage. The construction of such time-series datasets can provide a dynamic view of material properties; for example, observing a linear increase in permeability over time in the analysis can indicate a potential risk of seal failure, thus enabling effective monitoring in the field of materials science.

[0064] S105. Using a correlation analysis model, the time series dataset is processed to analyze the coupling relationship and hysteresis effect between the rate of chemical change of contents and the rate of physical degradation of shell under different environmental stress conditions.

[0065] For the initial associated itemset, conditional variables are used to group and classify different environmental stress conditions. This grouping is performed by classifying rate data using environmental stress condition thresholds, resulting in a rate index extraction result after grouping. This result extracts the slope of change as an index from the grouped data. From the extracted rate index results after grouping, an association analysis model is used to process the time series dataset. This model mines frequent itemsets using support and confidence, where support is the ratio of the itemset's occurrence frequency to the total number of transactions, and confidence is the conditional probability. This model determines the strength of the coupling relationship between the chemical change rate and the physical degradation rate, quantified based on the confidence value of frequent itemsets. Based on the strength of the coupling relationship, the lag effect delay is obtained. The delay difference under different environmental stress conditions is quantified through relationship pattern recognition, which calculates the peak position of the cross-correlation function from the intensity-related time series as the delay. From the delay difference, the effect prediction is determined. This prediction, after smoothing the delay data through data preprocessing steps, fits a regression curve to obtain the coupling relationship and lag effect between the chemical change rate of the contents and the physical degradation rate of the shell under different environmental stress conditions.

[0066] In one implementation, the time-series dataset originates from the field of material degradation monitoring, such as long-term test data of battery modules, containing time-series records of the rates of chemical change in the contents and the rates of physical degradation of the casing. The correlation analysis model first preprocesses the dataset, including data cleaning and normalization, to ensure data quality.

[0067] Specifically, the chemical change rate is calculated by monitoring the redox reaction rate of the contents, such as electrolytes, for example, based on the time derivative of concentration changes; the physical degradation rate is quantified by the decay of the elastic modulus of the shell material or the crack propagation rate. These rates are extracted as time-series features under different environmental stress conditions, such as high temperature, high humidity, or mechanical vibration. Furthermore, the association analysis model employs a variant of the Apriori algorithm, suitable for association rule mining of time series data.

[0068] For example, the model identifies frequent itemsets between chemical change rates and physical degradation rates; for instance, under high-temperature stress, an acceleration of the chemical rate may be accompanied by an increase in physical degradation. By calculating support and confidence, the model quantifies the coupling relationship, i.e., the strength of the mutual influence between the two rates.

[0069] It should be noted that the coupling relationship refers to how chemical changes affect the physical structure of the shell through thermal effects or corrosion processes, and vice versa.

[0070] In one possible implementation, the Pearson correlation coefficient is used to help verify the coupling strength and ensure the reliability of the analysis.

[0071] Preferably, a time window mechanism is introduced into the model to analyze the lag effect.

[0072] Specifically, a sliding window is used to segment the time series, for example, with a window size of 24 hours, to calculate the impact of the chemical change rate in the previous window on the physical degradation rate in the subsequent window. The lag time is quantified using a cross-correlation function; for example, it was found that under high-temperature conditions, chemical change leads physical degradation by approximately 12 hours. This lag effect reveals the dynamic evolution of the degradation process and helps predict material failure.

[0073] In one embodiment, for high-humidity environmental stress, the dataset includes humidity levels ranging from 60% to 90%. A correlation analysis model processes this data, first extracting chemical change rates, such as the degradation rate due to moisture absorption by the contents, and then correlating them with the physical degradation rate of the outer shell, such as the rate of increase in corrosion depth. For example, the model generates a rule such as "chemical rate > 0.5 units / hour ⇒ physical rate increases by 20%", with a support of 0.7 and a confidence level of 0.85. In this way, the coupling relationship is reflected in humidity-induced chemical penetration accelerating shell degradation, while the hysteresis effect manifests as physical degradation appearing approximately 6 hours after the chemical change. This embodiment demonstrates the application of the model under humid conditions, enabling a deeper understanding of the degradation mechanism.

[0074] Understandably, the model is further extended under mechanical vibration stress conditions. Time-series datasets capture the impact of vibration frequency on the chemical stability of the contents and the physical integrity of the shell. Correlation analysis, through frequent pattern mining, analyzes coupling relationships, such as vibration-induced microcracks promoting chemical leakage, thereby amplifying the degradation rate. The hysteresis effect is assessed using Granger causality tests, confirming that physical degradation lags behind chemical changes by approximately 4 hours. This approach enhances the versatility of the technical solution, making it suitable for material monitoring under similar vibration environments.

[0075] Specifically, the overall process begins with data acquisition, followed by the application of correlation analysis models to process time series data, and finally outputs quantitative results of coupling relationships and lag effects.

[0076] For example, in battery casing degradation scenarios, this analysis can identify critical stress thresholds to prevent sudden failures. Through objective rule extraction and statistical validation, this approach provides a reliable basis for degradation prediction.

[0077] S106. If the correlation analysis model identifies that the physical degradation of the outer shell has a significant promoting effect on the chemical changes of the contents, then the environmental stress condition is determined to be the critical acceleration path for product failure.

[0078] For the aforementioned environmental stress conditions, degradation rate coupling data are obtained from a time-series dataset. A correlation analysis model is used to calculate support and confidence levels. Support is the ratio of itemset occurrence frequency to total transactions, and confidence is a conditional probability, yielding a significant facilitator index. Based on this index, the intensity of the physical degradation of the outer shell on the chemical changes of the contents is determined. If the intensity exceeds a preset threshold, a key acceleration judgment result is established. From the key acceleration judgment result, the hysteresis delay is obtained. The delay difference is quantified using the peak position of a cross-correlation function. The input of this function is an intensity-related time series, and the output is the peak position as the delay, yielding degradation threshold monitoring parameters. For these parameters, conditional variable grouping is used to segment product failure paths, resulting in extended analysis results of coupling relationships and hysteresis effects.

[0079] In one implementation, a correlation analysis model is used to handle the relationship between data on the physical degradation of the product casing and data on the chemical changes of the contents.

[0080] Specifically, the model first collects test data on the product under different environmental stress conditions, such as shell deformation indices and changes in the pH value of the contents under factors like temperature, humidity, or vibration. Through statistical correlation analysis, the model calculates the Pearson correlation coefficient between physical degradation parameters, such as the shell thickness reduction rate, and chemical change parameters, such as the degree of oxidation. If the coefficient exceeds a preset threshold of 0.7, a significant promoting effect is identified. This process ensures data-driven objective judgment and avoids subjective interference. Furthermore, the correlation analysis model is constructed based on a machine learning framework, such as using linear regression to fit the functional relationship between degradation and change.

[0081] For example, in failure analysis of products such as lithium batteries, physical degradation of the casing manifests as a reduction in thickness due to surface corrosion, while chemical changes in the contents are caused by the decomposition of the electrolyte to generate gases. By inputting a historical test dataset into the model, the training process includes feature extraction and parameter optimization, and the output predictive model is used for real-time monitoring. This framework allows the model to adapt to various environmental stress scenarios while remaining within the scope of battery products.

[0082] It should be noted that the promoting effect of shell physical degradation on the chemical changes of the contents was identified through a causal reasoning mechanism. The specific process involves constructing a causal graph, where shell degradation is the dependent variable, chemical change is the effect variable, and environmental stresses such as high temperature are the intervention factors. The model uses a do-calculus method to simulate the intervention effect; if the promoting effect weakens after removing environmental stresses, its role as an accelerating pathway is confirmed.

[0083] For example, in one battery test embodiment, model analysis showed that high-temperature casing corrosion accelerated electrolyte degradation by up to 30%, thus identifying high temperature as the critical path. This explanation clarifies the causal analysis logic of the model, ensuring the traceability of business processes.

[0084] Preferably, in another implementation, the model integrates multivariate analysis to enhance robustness.

[0085] Specifically, time-series data, such as the rate of degradation of the outer casing over time, is introduced and combined with dynamic monitoring of the chemical indicators of the contents. Future trends are predicted using an ARIMA model; if the prediction shows that degradation significantly promotes change, a decision is triggered. This method is widely used in battery aging testing and can achieve early failure warnings.

[0086] For example, in an embodiment of a sealed container product, the association analysis model addresses the impact of physical degradation of the outer shell, such as material fatigue, on internal chemical reactions. The process includes data preprocessing to remove noise, followed by the application of association rule mining algorithms, such as the Apriori method, to extract rules such as "outer shell cracks ⇒ increased oxidation of contents." If the support exceeds 0.5 and the confidence level is higher than 0.8, a facilitating effect is identified, and vibration stress is determined to be a critical accelerating path. This example demonstrates the model's versatility in similar fields.

[0087] In one possible implementation, the model's output includes a visualization report showing the intensity of the promoting effect. This diagram allows users to intuitively understand the impact path of environmental stresses.

[0088] For example, intensity maps represent the distribution of correlation coefficients in the form of heat maps, which facilitates business decision-making.

[0089] Understandably, this technical solution, through the implementation of a correlation analysis model, can provide accurate path determination in product failure assessment. When applied on battery production lines, this method helps optimize design and reduce the risk of environmental stress-induced failures.

[0090] Specifically, this is further extended to a data acquisition implementation method integrating sensors. Sensors monitor the strain of the outer shell and changes in the spectral lines of the contents in real time, and the data is fed into the model for correlation calculations. If the model identifies a significant p-value less than 0.05 for degradation-induced changes, the corresponding environmental condition is determined to be a critical path. This integration method improves the real-time performance of the analysis.

[0091] In one embodiment, the model also considers the interactive effects of multiple environmental stresses.

[0092] For example, when temperature and humidity are applied simultaneously, the contribution of the interaction term is calculated. If the contribution rate exceeds twice that of a single stress, it is prioritized as an acceleration path. This description highlights the complex business scenarios handled by the model. Finally, in the effect description, this implementation achieves objective identification of product failure paths and can effectively guide stress screening strategies in battery reliability testing.

[0093] S107. Based on the identified key acceleration paths and corresponding environmental stress conditions, the initial shelf life prediction model is modified, wherein the modification includes introducing the shell state as an independent variable in the content degradation kinetic equation.

[0094] Key acceleration paths and corresponding environmental stress conditions are obtained from a pre-set environmental database to determine the path-stress correspondence. Based on this path-stress correspondence, shell state data is acquired and used as the independent variable in the content degradation kinetic equation to revise the initial shelf-life prediction model. This content degradation kinetic equation takes time and stress as inputs and outputs a degradation curve. Through this revised model, a packaging integrity assessment is introduced. This assessment obtains an integrity score by scanning the shell surface, and the model parameters are adjusted to obtain an estimated degradation rate. Based on this estimated degradation rate, the prediction accuracy is optimized, improving the shelf-life prediction accuracy.

[0095] In one implementation, the initial shelf-life prediction model is built upon the basic degradation rate of the product contents, for example, by describing the effect of temperature on degradation using the Arrhenius equation. Critical acceleration pathways refer to combinations of environmental factors that significantly accelerate the degradation of the contents, such as the high-temperature-high-humidity pathway, which are identified through accelerated aging tests.

[0096] Specifically, the process begins by collecting experimental data on the product under various environmental conditions, including stress conditions such as temperature, humidity, and light exposure. Then, it analyzes which combinations lead to a sharp increase in degradation rate, thereby identifying critical pathways. Furthermore, the corresponding environmental stress conditions include quantitative indicators, such as a temperature range of 20-40 degrees Celsius and a relative humidity of 60-90%. These conditions are monitored in real time by sensors and used as input parameters in model correction. The correction process begins by evaluating the prediction bias of the initial model, for example, comparing the predicted shelf life with the actual aging test results. If the bias exceeds a threshold, a correction mechanism is introduced.

[0097] Preferably, the shell state is introduced as an independent variable in the content degradation kinetic equation. The shell state refers to the integrity of the packaging material, such as changes in its seal or permeability. In practice, the shell state can be quantified through optical scanning or permeability testing, for example, by measuring the microcrack density or oxygen permeability of the shell and converting it into a numerical variable. The original content degradation kinetic equation might be dC / dt = -k*C, where C is the content concentration and k is the rate constant. After modification, the shell state S is introduced as an independent variable, becoming dC / dt = -k(S)*C, where k(S) is a function dependent on S, such as k(S) = k0 * exp(a*S), where a is an empirical coefficient. Through this introduction, the shell state directly affects the degradation rate. For example, if shell damage leads to an increase in S, then k(S) increases, predicting a shorter shelf life.

[0098] In one possible implementation, for food applications such as canned beverages, the key acceleration pathway might be a combination of light exposure and oxygen permeation. Environmental stress conditions, including UV intensity and oxygen concentration, are first identified through laboratory simulations, such as placing samples in controlled environments and monitoring pH or color changes. When refining the model, the shell condition, such as the cap seal, is incorporated into the equation, and the independent variable S can be adjusted from an initial value of 0 (intact) to 1 (damaged), thereby adjusting the degradation kinetics and ensuring more accurate predictions.

[0099] It should be noted that this modification enhances the model's adaptability to real-world storage conditions.

[0100] For example, in predicting the shelf life of pharmaceuticals, the critical path is the interaction between humidity and temperature. After quantifying the stress conditions, the humidity permeability of the bottle shell is introduced as an independent variable, and the modified equation can reflect the impact of packaging aging on drug stability.

[0101] Specifically, in the implementation process, an initial model is first established, and then a correction factor is calculated based on the identified path. For example, for cosmetic products, the acceleration path is a combination of thermal stress and vibration. The stress condition is such as a vibration frequency of 10-50Hz. During correction, the shell condition, such as changes in container wall thickness, is introduced, and the kinetic equation is adjusted to consider the degradation rate with enhanced penetration. Furthermore...

[0102] In one embodiment, model correction can be achieved through software tools, such as iteratively optimizing parameters using numerical simulation software. After environmental stress data is input, the system automatically incorporates shell state variables and outputs a corrected shelf-life prediction curve.

[0103] For example, in the field of battery storage, the critical path for shelf life prediction is high temperature and electrolyte leakage. Corrections include taking the casing corrosion state as an independent variable, which affects the electrolyte degradation equation.

[0104] Understandably, this approach is applicable to multiple scenarios within the same field, such as different types of food packaging, ensuring versatility. After implementation, this correction enables a more accurate response to environmental variations; for example, in testing, the predicted error of the corrected model was reduced to 50% of that of the initial model.

[0105] S108. Using the modified shelf life prediction model, input the typical environmental stress spectrum of different market regions, and calculate the differentiated product shelf life prediction values ​​for each regional market.

[0106] A typical environmental stress spectrum for the market region is obtained, and temperature and humidity distribution data are extracted from a pre-set database to determine the quantitative indicators of the stress spectrum. These quantitative indicators are then input into a modified shelf-life prediction model. An accelerated life testing method is used to adjust the model parameters to account for the effects of temperature and humidity. This accelerated life testing method applies stress higher than normal levels to accelerate product degradation and fits the lifespan distribution, resulting in a region-specific correction coefficient. The degradation rate of the product in each region is calculated based on this correction coefficient, obtained by multiplying the correction coefficient by a standard degradation function. If the degradation rate exceeds a preset threshold, the initial prediction value is reduced to determine the differentiated shelf-life range. Predicted values ​​for each market are output from these differentiated shelf-life ranges. These values ​​are then combined with supply chain inventory data obtained from the pre-set database to generate adjusted quotas, resulting in the final regional product shelf-life prediction value.

[0107] In one implementation, the modified shelf-life prediction model is optimized based on historical data and environmental factors to assess the shelf life of a product under different conditions.

[0108] Specifically, the model first integrates the product’s basic attributes, such as ingredients and packaging type, and then introduces correction parameters to adjust the initial predictions.

[0109] For example, for food products, the model can consider the impact of temperature fluctuations on microbial growth and improve accuracy through algorithmic adjustments. The environmental stress spectrum refers to the distribution spectrum of typical combinations of environmental factors within a specific market area, including variables such as temperature, humidity, light intensity, and atmospheric pressure.

[0110] It should be noted that the process of constructing the environmental stress spectrum involves collecting historical meteorological data of the region and generating a representative spectrum through statistical analysis.

[0111] For example, in tropical market regions, the environmental stress spectrum might emphasize the frequency distribution of high humidity and high temperature, while in cold regions it would highlight the risk of freezing at low temperatures. This spectrum is represented in vector form, with each dimension corresponding to a statistical characteristic value of an environmental variable, thus providing the input basis for the model. Furthermore, the implementation of the modified shelf-life prediction model involves several key steps. First, the initial model is modified, which might be based on the Arrhenius equation simulating chemical reaction rates, but the modification process introduces region-specific coefficients to adjust the parameters.

[0112] Specifically, the correction algorithm calculates the impact of environmental stress on product degradation rates, for example, by integrating phylogenetic data using a weighted averaging method to generate a correction factor. This factor is applied to the initial prediction formula, outputting the adjusted shelf-life value. This method ensures that the model adapts to the actual conditions of different regions, improving the reliability of the prediction.

[0113] In one possible implementation, for the input process in different market regions, a typical environmental stress spectrum is first obtained.

[0114] For example, in the Asian market, spectral data is sourced from meteorological databases, covering seasonal humidity variations; while in the European market, the focus is on light intensity distribution. The model uses these spectral data as input vectors to calculate the product degradation curve for each region. Specifically, the model simulates the cumulative effect of environmental stress on product stability, estimating total stress exposure through integral methods to derive differentiated shelf-life predictions, such as a shortened shelf life of 6 months in tropical regions and an extended shelf life of 12 months in cold regions.

[0115] Preferably, the forecasting process can be extended to various product types, such as dairy products or electronic components, but is limited to the field of consumer product shelf-life management.

[0116] For example, in the dairy industry, the model focuses on analyzing the amplifying effect of humidity on bacterial growth, calculating regionalized predictions; in the electronic component scenario, it examines the impact of temperature on material aging. This diverse application demonstrates the model's versatility without altering its core business areas.

[0117] For example, in actual operation, after the user inputs the environmental stress spectrum of a specific area, the model automatically performs the calculation.

[0118] In one embodiment, for the North American market, the spectrum includes peak winter temperatures, which the model adjusts its predictions to output a 9-month shelf life. This process is implemented through a modular design, including a data input module, a correction calculation module, and an output module, ensuring the independence and traceability of each step.

[0119] Understandably, this differentiated forecasting helps optimize supply chain management, such as adjusting product distribution strategies to match regional shelf life. Through this approach, companies can reduce waste and improve product safety, while the model's correction mechanism ensures the accuracy of the forecasts. Furthermore, in another embodiment, the model integrates real-time data updates.

[0120] For example, when the regional environmental stress spectrum changes, such as due to anomalous climate events, the model recalculates the predicted values. This dynamic adjustment enhances the flexibility of the scheme and supports long-term application.

[0121] Specifically, the core of calculating the predicted shelf life of differentiated products lies in the coupling process between the environmental stress spectrum and the model. This process first quantifies the weight of each variable in the spectrum, for example, temperature has a higher weight than humidity, and then integrates them into the degradation model through a product. The result output is a region-specific list of values ​​for decision-making. This detailed calculation ensures the scientific validity and practicality of the prediction.

[0122] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for assessing and predicting the environmental adaptability of soft capsules, characterized in that, include: The process involves: acquiring environmental data and logistics / warehousing scenario data for the target product's intended sales area; generating an environmental risk sequence based on the environmental and logistics / warehousing scenario data; obtaining product tolerance analysis results through simulation analysis; predicting the potential damage probability using the product tolerance analysis results; generating a composite environmental stress spectrum with time-series characteristics, which simulates the environmental pressure combination of the product from production to the consumer end; conducting accelerated aging experiments on product samples based on the composite environmental stress spectrum, collecting aging-related data, and determining stability indicators; constructing a time-series dataset and analyzing the coupling relationships and hysteresis effects among different indicators in the time-series dataset. Based on the analysis results, key acceleration paths are identified, and the shelf life prediction model is revised. Using the revised shelf life prediction model, the environmental stress spectrum of different market regions is input to calculate differentiated shelf life prediction values.

2. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The acquisition of environmental data and logistics warehousing scenario data for the target product's intended sales area includes: extracting historical climate data for the intended sales area from a preset meteorological database to obtain temperature and humidity sequences and light intensity sequences. The temperature and humidity sequences record daily averages, and the light intensity sequences measure cumulative irradiation duration using sensors. Collecting typical logistics warehousing scenario data through IoT sensors to determine mechanical vibration frequency and amplitude. The mechanical vibration frequency is recorded using accelerometers to capture peak fluctuations during transportation, and the amplitude is obtained through frequency domain signal conversion. Integrating the temperature and humidity sequences and the mechanical vibration frequency data to generate the environmental risk sequence. If the environmental risk sequence exceeds a preset threshold, adjusting warehousing parameters. The data integration calculates a composite value using a weighted average, where the weighted average is the sum of each sequence value multiplied by a preset weight. Simulating a transportation impact assessment based on the environmental risk sequence yields the product tolerance analysis results. The product tolerance analysis results generate distribution characteristics through repeated random sampling, with the environmental risk sequence as input and the distribution characteristics as output.

3. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The generation of a composite environmental stress spectrum with time-series characteristics, which simulates the environmental pressure combination that a product may experience from production to consumption, includes: acquiring multi-dimensional environmental data, extracting temperature, humidity, and vibration indicators from production process pressures and end-consumer experiences to form an initial dataset; performing time-series feature analysis on the initial dataset using environmental simulation algorithms, determining the data dimension analysis results by extracting sequence patterns and correlation indicators; generating a stress spectrum generation sequence based on the data dimension analysis results, integrating multi-dimensional simulation elements; and constructing a pressure combination generation model for the stress spectrum generation sequence, outputting the composite environmental stress spectrum with time-series characteristics by integrating sequence elements and pressure indicators. The composite environmental stress spectrum characterizes the environmental pressure combination that a product may experience from production to consumption.

4. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The accelerated aging experiment on the product sample based on the composite environmental stress spectrum, collecting aging-related data and determining stability indicators includes: acquiring the composite environmental stress spectrum, conducting an accelerated aging experiment on the product sample in a controlled experimental chamber using the composite environmental stress spectrum, and obtaining aging process data; simultaneously collecting the concentration data of key chemical components of the contents and the physical performance parameter data of the shell using the aging process data; comparing the concentration and parameter deviations between samples from the collected data to determine batch sample differences; for the batch sample differences, using a preset threshold to judge the concentration change trend and performance degradation degree, if the deviation exceeds the preset threshold, it is marked as a significant change, and an integrated dataset is obtained; fitting an aging curve from the integrated dataset, determining the curve equation through data point processing, and obtaining the stability indicators.

5. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The construction of the time-series dataset and the analysis of the coupling relationships and lag effects among different indicators in the time-series dataset include: acquiring the content of specific degradation products of the contents and the hardness and permeability coefficient of the shell through acquisition devices to form initial time-series data; integrating chemical change indicators and physical degradation indicators using a weighted average fusion method based on the initial time-series data to obtain a standardized dataset; if the standardized dataset shows correlation between indicators, determining the degradation trend monitoring value through correlation calculation; acquiring the degradation trend monitoring value, identifying outliers, and optimizing through interpolation to obtain a complete time-series dataset; and using an association analysis model to process the complete time-series dataset and analyze the coupling strength between the chemical change rate and the physical degradation rate under different environmental stress conditions, wherein the coupling strength is quantified based on the confidence value of frequent itemsets.

6. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The construction of the time series dataset and the analysis of the coupling relationships and lag effects among different indicators in the time series dataset include: For the initial associated itemset, conditional variable grouping is used to classify different environmental stress conditions. The grouping is performed by classifying rate data through environmental stress condition thresholds to obtain the rate indicator extraction results after grouping. The slope of change is extracted from the grouped data as an indicator. From the rate indicator extraction results after grouping, the time series dataset is processed using an association analysis model. Frequent itemsets are mined through support and confidence to determine the strength of the coupling relationship between the chemical change rate and the physical degradation rate. Based on the strength of the coupling relationship, the lag effect delay is obtained. The delay difference under different environmental stress conditions is quantified by relational pattern recognition. The recognition is performed by calculating the peak position of the cross-correlation function from the intensity-related time series as the delay. From the delay difference, the effect impact prediction is determined. The prediction is performed by fitting a regression curve after smoothing the delay data to obtain the coupling relationship and lag effect.

7. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The identification of key acceleration paths based on the analysis results includes: For environmental stress conditions, obtaining degradation rate coupling data from the time series dataset, calculating support and confidence using an association analysis model to obtain a significant facilitator index, where support is the frequency of itemset occurrence and the proportion of total transactions, and confidence is a conditional probability; Based on the significant facilitator index, determining the intensity of the physical degradation of the shell on the chemical changes of the contents; if the intensity exceeds a preset threshold, determining a key acceleration result; From the key acceleration result, obtaining the hysteresis delay, quantifying the delay difference through the peak position of the cross-correlation function, and obtaining degradation threshold monitoring parameters; For the degradation threshold monitoring parameters, using conditional variable grouping to divide the product failure paths, obtaining extended analysis results of coupling relationships and hysteresis effects, which are used for subsequent model correction.

8. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The revised shelf-life prediction model includes: obtaining the key acceleration path and corresponding environmental stress conditions from a preset environmental database, and determining the path-stress correspondence; obtaining shell state data based on the path-stress correspondence, using the shell state data as the independent variable of the content degradation kinetic equation, and revising the initial shelf-life prediction model, wherein the content degradation kinetic equation takes time and stress as input and outputs a degradation curve; introducing packaging integrity assessment through the revised shelf-life prediction model, wherein the packaging integrity assessment obtains an integrity score by scanning the shell surface, adjusting model parameters, and obtaining an estimated degradation rate; optimizing the prediction accuracy based on the estimated degradation rate, and generating shelf-life prediction results applicable to different environmental conditions.

9. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The step of calculating differentiated shelf-life prediction values ​​by inputting environmental stress spectra of different market regions into the modified shelf-life prediction model includes: acquiring typical environmental stress spectra of market regions, extracting temperature and humidity distribution data from a preset database, and determining quantitative indicators of the environmental stress spectra; inputting the quantitative indicators into the modified shelf-life prediction model, adjusting model parameters for the effects of temperature and humidity using accelerated life testing methods to obtain region-specific correction coefficients, wherein the accelerated life testing method accelerates product degradation and fits the life distribution by applying stress higher than normal levels; calculating the product degradation rate in each region based on the correction coefficients, wherein the degradation rate is obtained by multiplying the correction coefficients by a standard degradation function, and if the degradation rate exceeds a preset threshold, reducing the initial prediction value to determine the differentiated shelf-life range; outputting the prediction values ​​for each market from the differentiated shelf-life range, generating adjustment quotas by combining supply chain inventory data, and obtaining the final regional product shelf-life prediction values.

10. The method for assessing and predicting the environmental adaptability of soft capsules as described in claim 1, characterized in that, The step of using the product tolerance analysis results to predict the potential damage probability and determine the storage suitability of the target product in the intended sales area includes: extracting distribution characteristic data based on the product tolerance analysis results, wherein the distribution characteristic data is generated through simulation analysis and characterizes the product's tolerance under different environmental conditions; calculating the potential damage probability based on the distribution characteristic data, wherein the potential damage probability is extracted from the distribution characteristic data through statistical analysis methods; adjusting storage condition parameters if the potential damage probability exceeds a preset threshold, wherein the storage condition parameters include temperature control range and humidity control range; and determining the storage suitability of the target product in the intended sales area based on the potential damage probability and the adjusted storage condition parameters, wherein the storage suitability is obtained by weighted calculation of comprehensive environmental risk and tolerance, and is used to guide the formulation of subsequent warehousing and transportation strategies.

Images