Intelligent packaging material selection method and system based on multi-dimensional data analysis

Abstract

The invention provides a packaging material intelligent selection method and system based on multi-dimensional data analysis, and relates to the field of material management. Vibration, deformation and temperature and humidity data are collected in real time through an environment sensing module to construct a feature matrix; the data analysis module fuses the historical case library and the logistics task duration to predict the material life, and associates the supply chain inventory to generate residual life evaluation; the decision engine module dynamically distributes cost, safety and environmental protection weights, and combines the supplier credit score to generate a multi-level decision scheme; the execution control module verifies the dimensional tolerance of the material and drives replacement operation, and model calibration is triggered through a 5% deviation threshold value; a humidity sensitive material grading standard and a deformation safety limit value are built in the system, and when the temperature and humidity sudden change exceeds the threshold value, deformation quantity prediction and emergency scheme reconstruction are started; a three-level self-optimization mechanism of environmental perception error, material life error and execution error is established, and closed-loop optimization is realized through sensor calibration, supplier credit recalculation and dynamic adjustment of control parameters.

Description

Technical Field

[0001] The present invention relates to the field of material management, and specifically to an intelligent selection method and system for packaging materials based on multi-dimensional data analysis. Background Technique

[0002] Modern logistics transportation places higher requirements on the performance of packaging materials; the globalization of commodity circulation intensifies the complexity of transportation environments; traditional material selection methods rely on manual experience and static parameters; vibrations, temperature, and humidity changes during transportation lead to dynamic attenuation of material properties; the upgrading of environmental protection regulations requires considering both material costs and sustainability; intelligent decision-making systems have become a key direction for industry upgrading.

[0003] Current mainstream solutions use single environmental sensors to monitor transportation conditions; establish static selection models based on material hardness and thickness; some systems introduce historical transportation data to establish linear prediction formulas; a few solutions integrate supplier databases for inventory matching; use fixed-weight algorithms to balance cost and safety indicators; trigger manual intervention processes through threshold alarm mechanisms.

[0004] The deficiencies of the existing technology are as follows: lack of multi-source data fusion and analysis capabilities; single monitoring dimension of environmental parameters leads to prediction deviations; static models cannot reflect the dynamic attenuation law of material properties; inventory matching ignores supplier production capacity fluctuations and logistics timeliness; fixed-weight algorithms are difficult to adapt to changes in transportation stage requirements; delayed response of manual intervention causes risks of lagging material replacement. Summary of the Invention

[0005] (I) Technical Problems to be Solved

[0006] In view of the deficiencies of the existing technology, the present invention provides an intelligent selection method and system for packaging materials based on multi-dimensional data analysis to solve the problems of lack of multi-source data fusion and analysis capabilities; single monitoring dimension of environmental parameters leading to prediction deviations; static models unable to reflect the dynamic attenuation law of material properties; inventory matching ignoring supplier production capacity fluctuations and logistics timeliness; fixed-weight algorithms being difficult to adapt to changes in transportation stage requirements; and delayed response of manual intervention causing risks of lagging material replacement as mentioned in the above background technique.

[0007] (II) Technical Solutions

[0008] To achieve the above objectives, the present invention is realized through the following technical solutions: an intelligent selection method and system for packaging materials based on multi-dimensional data analysis, including an environmental perception module, a data analysis module, a decision engine module, and an execution control module;

[0009] The environmental perception module uses a sensor network deployed on the surface of the packaging box to collect vibration spectrum, surface strain, and temperature and humidity data in real time, and generates an environmental state matrix after feature extraction. The sensor network includes an automatic calibration unit;

[0010] The data analysis module has a built-in historical transportation case database that stores the performance decay records of different materials in the transportation environment. After receiving the environmental status matrix, it establishes an association model between the vibration energy distribution and the material performance decay, predicts the remaining life of the material in combination with the inventory turnover rate data in the supply chain database, and establishes a data interface with the logistics management system to obtain the transportation task duration and route planning data in real time;

[0011] The decision engine module dynamically adjusts the priorities of cost, safety, and environmental protection goals according to the transportation stage, fuses the material life prediction results with the supplier production capacity data, and generates a multi-level decision tree containing automatic execution plans and emergency plans;

[0012] The execution control module pre-stores the packing box size tolerance standard, the structural deformation safety limit, and the registered size database; and converts the selected plan into a material replacement instruction. After driving the actuator to complete the operation, it collects the error between the actual transportation data and the predicted value and feeds it back to the data analysis module to trigger the dynamic calibration of the model parameters. The model prediction errors are divided into three categories: environmental perception error, material performance error, and execution control error;

[0013] The supply chain database includes a material moisture-proof characteristic index table and a hygroscopic expansion coefficient safety critical value library, records the deformation coefficients, protection levels, and irreversible deformation thresholds of each material in different humidity environments, and marks the humidity sensitivity level; at the same time, it dynamically calculates based on the delivery on-time rate and the quality inspection pass rate, constructs a supplier credit scoring system, and sets the credit scoring threshold as the value of the normal distribution curve of historical performance data.

[0014] Preferably, the environmental perception module realizes the collection of transportation environment data through a sensor network deployed on six sides of the packing box; the sensor network includes three types of detection units: vibration sensors, deformation sensors, and temperature and humidity sensors; the vibration sensors collect three-dimensional acceleration data at a sampling rate of 1 kHz, and extract the energy distribution characteristics of the 0-500 Hz frequency band through fast Fourier transform; the deformation sensors monitor the surface micro-strain at a frequency of 100 Hz, and use a spatial interpolation algorithm to construct a deformation gradient field model; the temperature and humidity sensors synchronously detect the environmental parameters of each side, and trigger an abnormal mark when the temperature difference between adjacent sides exceeds 3°C or the humidity difference exceeds 15%RH; all sensor data is transmitted to the edge node through a low-power wireless network, and timestamp alignment and standardization processing are performed to generate an environmental status matrix containing the frequency-domain energy spectrum, deformation gradient, and temperature and humidity distribution.

[0015] Preferably, the data analysis module has a built-in historical transportation case database that stores the performance degradation records of materials under the combined action of vibration, temperature, and humidity. After receiving the environmental state matrix, it extracts the characteristic of the proportion of the main vibration frequency energy, and conducts a matching degree analysis with the material natural frequency database. When it detects that the proportion of energy in the resonance frequency band exceeds the preset threshold, it starts the real-time re-evaluation process of the material buffering efficiency. Combining the inventory turnover rate data in the supply chain database, it calculates the confidence interval of the remaining life of the material. If the predicted life is lower than the transportation task duration obtained from the logistics management system, it triggers the real-time inventory verification process.

[0016] Preferably, the decision engine module dynamically adjusts the weights of the objective function according to the transportation stage. In the transportation preparation stage, the cost weight is set to 0.6, the safety weight is set to 0.3, and the environmental protection weight is set to 0.1. In the middle of transportation, the safety weight is increased to 0.5. In the customs clearance stage, environmental protection compliance constraints are introduced, and a 1.2-fold gain coefficient is applied to the environmental protection target. The module fuses the material life prediction result with the real-time production capacity data of the supplier to generate a three-level recommendation plan. When the inventory is less than 120% of the current transportation demand, it automatically reduces the priority of this material, and associates with the supplier database to screen a list of alternative suppliers with sufficient production capacity and logistics timeliness meeting the shortest delivery cycle.

[0017] Preferably, the execution control module pre-stores the packaging box size tolerance standard and the structural deformation safety limit value. After receiving the material specification parameters issued by the decision engine module, it matches and verifies the material size with the three-dimensional model of the packaging box in the registered size database. If it detects that the tolerance exceeds the preset safety range, it triggers the process of regenerating the plan based on the actual size constraint. It drives the pneumatic-electromagnetic hybrid actuator to complete the material replacement operation, and real-time feedbacks the positioning accuracy through the displacement sensor. When the measured position deviation exceeds 0.5 mm, it starts the PID dynamic deviation correction mechanism. After the operation is completed, it collects the actual transportation data, calculates the deviation from the predicted value, and feeds it back to the data analysis module.

[0018] Preferably, the supply chain database includes a material moisture-proof characteristic index table and a supplier credit scoring system. The moisture-proof characteristic index table records the deformation coefficient and protection level of each material in the humidity range of 30% - 90% RH. Materials with a moisture absorption expansion coefficient ≥ 0.5 mm / %RH are marked as humidity-sensitive. The credit scoring system is dynamically calculated based on the on-time delivery rate weight and the quality inspection pass rate weight. Among them, the on-time delivery rate weight is 0.6, and the quality inspection pass rate weight is 0.4. The credit scoring threshold is the μ - 2σ value of the normal distribution of historical performance data. When the supplier score is lower than the threshold, its material options are automatically blocked. The actual loss rate data fed back by the execution control module is included in the next cycle's score calculation with a weight of 0.2.

[0019] Preferably, when the energy value in the frequency band of the material's natural frequency ±10% identified by the vibration spectrum analysis increases by more than 50% of the baseline value within 2 seconds, the environmental perception module sends a resonance warning signal to the data analysis module; the data analysis module retrieves the transportation records with the same frequency band characteristics in the historical case library and counts the breakage rate of the corresponding materials; if the breakage rate exceeds the safety threshold, a request for re-evaluating the plan is initiated to the decision engine module; the decision engine module generates a priority list of alternative materials and associates with the supplier database to screen alternative suppliers that meet the delivery timeliness.

[0020] Preferably, the implementation process of the humidity-sensitive material processing strategy of the data analysis module is as follows: retrieve the list of materials with a hygroscopic expansion coefficient ≥ 0.5 mm / %RH from the supply chain database, and when the humidity change rate > 15%RH / h is found in the environmental perception data, start the dedicated prediction model; the prediction model uses an LSTM neural network, the input layer includes the current humidity value, the change gradient and the characteristics of the material's hygroscopic curve, and the output layer predicts the deformation amount within the next 2 hours; when the predicted deformation exceeds 80% of the structural deformation safety limit, trigger a three-level response mechanism: the first level sends a suggestion of downgrading use to the decision engine, the second level starts the automatic desiccant dispensing system, and the third level calls the moisture-proof material replacement plan; at the same time, mark the abnormal data packet as a high-value training sample and give priority to model iteration and update.

[0021] Preferably, the fast approval process of the execution control module is as follows: after the decision engine generates a recommended list of alternative materials, the system automatically retrieves the customs HS code database to verify the compliance of the materials, and adds an electronic fence mark to the plan involving restricted substances; push the approval request to the responsible engineer through the enterprise WeChat API, attaching a material parameter comparison table and a risk analysis report; the engineer uses a digital certificate for electronic signature confirmation, and the system automatically records the signing timestamp and device fingerprint; immediately issue a production work order to the MES system after the approval is passed, and synchronously update the WMS inventory status; the entire process needs to be completed within 15 minutes, and if the approval is not completed within the time limit, it will be automatically escalated to the superior supervisor and the preparatory emergency plan will be started.

[0022] Preferably, the operation of the model parameter dynamic calibration process includes: when receiving the calibration request from the execution control module, extract the original sensor data of the corresponding time window from the data lake; perform a joint time-frequency domain analysis on the vibration signal, and use the Wigner-Ville distribution to extract the instantaneous frequency characteristics; perform a residual analysis on the actual material loss data and the predicted value to locate that the error mainly comes from the dimension; for humidity-sensitive materials, preferably use the Bayesian optimization algorithm to update the humidity coupling coefficient in the attenuation model, and calculate the Jacobian matrix for each iteration to determine the parameter adjustment direction; after calibration, generate a model file with a version number identifier, and publish it to the decision engine and the production database after verification by digital signature.

[0023] Preferably, the data analysis module regularly classifies and attributes the model prediction errors; at the end of each quarter, a global health assessment is initiated, and the model prediction errors are classified by source: when the environmental perception error exceeds 5%, channel quality analysis and node distribution optimization are performed on the sensor network; when the material performance error exceeds 8%, the material samples provided by the supplier are resampled for destructive testing, and the basic parameter database is updated; when the execution control error exceeds 2 mm, the transmission components of the driving mechanism are detected for accuracy by a laser interferometer, and the ball screw with wear exceeding the tolerance band is replaced; all optimization operations form a closed-loop control, and after each adjustment, verification tests for three complete transportation cycles are required to ensure that the improvement of system stability meets the preset KPI indicators.

[0024] Preferably, after the installation of the new material is completed, the execution control module monitors the initial buffer efficiency parameters in real time; when the deviation between the measured value and the model prediction value exceeds 5%, a calibration request is sent to the data analysis module; the request includes the environmental data timestamp and the material batch code; the data analysis module extracts the environmental feature data for the corresponding period and preferentially updates the performance degradation model of the humidity-sensitive material; after the updated model parameters are verified by digital signature, they are synchronized to the decision engine module.

[0025] Preferably, the registered size database of the packaging box includes full-size measurement of each batch of packaging boxes using a three-dimensional laser scanner to collect the spatial coordinates of no less than 2000 feature points; the ICP algorithm is used to register the point cloud data with the design drawings, and the statistical distribution characteristics of each size parameter are calculated; dynamic tolerance bands are set for the length, width, and height respectively, and the width of the tolerance band is ±(0.1% × nominal value + 0.5 mm); when it is detected that the size standard deviation of the packaging boxes in the same batch exceeds 50% of the tolerance band, a quality exception report is automatically generated and the supplier deduction process is triggered; the database is defragmented and the index is rebuilt monthly to ensure that the query response time is less than 50 ms.

[0026] (III) Advantageous Effects

[0027] The present invention provides an intelligent selection method and system for packaging materials based on multi-dimensional data analysis. It has the following advantageous effects:

[0028] 1. The present invention significantly improves the accuracy and adaptability of packaging material selection through a multi-source data fusion and dynamic decision-making mechanism; the environmental perception module captures vibration spectrum, surface strain, and temperature and humidity gradient data in real time, establishes a material performance attenuation prediction model in combination with the historical transportation case database, effectively identifies resonance risks and deformation critical states, and ensures transportation safety; the decision engine module dynamically adjusts the weight of cost, safety, and environmental protection goals based on the priority of the transportation stage, generates multi-level recommendation schemes and correlates with real-time supply chain data to achieve optimal resource utilization; the execution control module controls the material replacement error within 5% through a positioning and closed-loop feedback mechanism, and continuously improves the system prediction accuracy through dynamic calibration of model parameters; the multi-dimensional data collaboration mechanism forms a closed-loop optimization of environmental monitoring, material life prediction, and supply chain scheduling.

[0029] 2. The intelligent collaboration network constructed by the present invention greatly enhances the supply chain response efficiency and risk resistance ability; the supply chain database integrates the moisture-proof property index table of materials and the moisture absorption and expansion safety critical value library, combines with the dynamic calibration mechanism of humidity-sensitive materials, quickly generates alternative solutions when the temperature and humidity change suddenly, and urgently executes through a three-step approval process; the supplier credit scoring system dynamically screens high-quality suppliers based on the normal distribution threshold, combines with the closed-loop feedback mechanism of actual loss data; the self-optimization mechanism triggers sensor calibration, supplier data review, and control parameter optimization respectively by classifying and attributing model errors, improving the system iteration efficiency by 50% per month; the introduction of the registered dimension database and the tolerance compatibility screening logic completely eliminates the size matching error between the packing box and the material. Detailed implementation manners

[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] The embodiment of the present invention provides an intelligent selection method and system for packaging materials based on multi-dimensional data analysis. The specific implementation process is as follows: when the system starts, the environmental perception module collects transportation environment data in real time through the sensor network deployed on six sides of the packing box; the vibration sensor captures three-dimensional acceleration data at a sampling rate of 1 kHz and extracts the energy distribution characteristics in the 0-500 Hz frequency band through fast Fourier transform; the deformation sensor monitors the surface micro-strain at a frequency of 100 Hz and constructs a deformation gradient field model using a spatial interpolation algorithm; the temperature and humidity sensor synchronously detects the environmental parameters of each side, and triggers an abnormal mark when the temperature difference between adjacent sides exceeds 3°C or the humidity difference exceeds 15%RH.

[0032] All sensor data is transmitted to the edge node through a low-power wireless network, and timestamp alignment and normalization processing are performed to generate an environmental state matrix containing the frequency-domain energy spectrum, deformation gradient, and temperature and humidity distribution. After receiving the environmental state matrix, the data analysis module extracts the feature of the proportion of the main vibration frequency energy and analyzes the matching degree with the material natural frequency database. When it is detected that the proportion of the energy in the resonance frequency band exceeds the preset threshold, the real-time re-estimation process of the material buffer efficiency is started. The confidence interval of the remaining life of the material is calculated by combining the inventory turnover rate data in the supply chain database. If the predicted life is lower than the transportation task duration obtained from the logistics management system, the real-time inventory verification process is triggered.

[0033] Among them, the decision engine module dynamically adjusts the weights of the objective function according to the transportation stage. In the transportation preparation stage, the cost weight is set to 0.6, the safety weight is set to 0.3, and the environmental protection weight is set to 0.1. In the middle stage of transportation, the safety weight is increased to 0.5. In the customs clearance stage, environmental protection compliance constraints are introduced to apply a 1.2-fold gain coefficient to the environmental protection target. The module fuses the material life prediction results with the real-time production capacity data of the supplier to generate a three-level recommendation plan. When the inventory is less than 120% of the current transportation demand, the priority of this material is automatically reduced, and the supplier database is associated to screen a list of alternative suppliers with sufficient production capacity and logistics timeliness meeting the shortest delivery cycle.

[0034] The packaging box size tolerance standard and the structural deformation safety limit are pre-stored in the execution control module. After receiving the material specification parameters issued by the decision engine module, the material size is matched and verified with the three-dimensional model of the packaging box in the registered size database. If it is detected that the tolerance exceeds the preset safety range, the process of re-generating the plan based on the actual size constraint is triggered. The pneumatic-electromagnetic hybrid actuator is driven to complete the material replacement operation, and the positioning accuracy is fed back in real time through the displacement sensor. When the measured position deviation exceeds 0.5 mm, the PID dynamic deviation correction mechanism is started. After the operation is completed, the deviation between the actual transportation data and the predicted value is collected and fed back to the data analysis module to trigger the dynamic calibration of the model parameters.

[0035] The credit scoring system in the supply chain database is dynamically calculated based on the weight of 0.6 for on-time delivery rate and 0.4 for quality inspection pass rate. When the supplier score is lower than the μ-2σ value of the normal distribution of historical performance data, its material options are automatically blocked; the actual loss rate data fed back by the execution control module is included in the next cycle score calculation with a weight of 0.2, forming a closed-loop feedback mechanism; when the vibration spectrum analysis identifies that the energy value of the material's natural frequency ±10% frequency band increases by more than 50% of the baseline value within 2 seconds, the environmental perception module sends a resonance warning signal to the data analysis module, and the data analysis module retrieves the transportation records of the same frequency band characteristics in the historical case library, and calculates the breakage rate of the corresponding material. If the breakage rate exceeds the safety threshold, a request for re-evaluation of the plan is initiated to the decision engine module. The decision engine generates a priority list of alternative materials and links the supplier database to screen alternative suppliers that meet the delivery time limit. When the humidity-sensitive material processing strategy is implemented, the material list with a hygroscopic expansion coefficient ≥ 0.5mm / %RH is retrieved from the supply chain database. When the humidity change rate ≥ 15%RH / h is found in the environmental perception data, the dedicated LSTM neural network prediction model is started. The input layer contains the current humidity value, the change gradient and the material moisture absorption curve characteristics. The output layer predicts the deformation within the next 2 hours. When the predicted deformation exceeds 80% of the safety limit of the structural deformation, the third-level response mechanism is triggered.

[0036] In the fast approval process, the system automatically retrieves the customs HS code database to verify the material compliance, adds electronic fence marks to the scheme involving restricted substances, and pushes the approval request to the responsible engineer through the enterprise WeChat API, with a material parameter comparison table and risk analysis report. The engineer uses a digital certificate to confirm the electronic signature. The system automatically records the signature timestamp and device fingerprint. After approval, the production work order is immediately issued to the MES system, and the WMS inventory status is updated simultaneously. The entire process must be completed within 15 minutes. If it is not approved within the time limit, it will be automatically escalated to the superior supervisor and the emergency plan will be initiated. At the end of the quarter, the global health assessment is initiated, and the model prediction error is classified by source. When the environmental perception error exceeds 5%, the sensor network is analyzed for channel quality. When the material performance error with node distribution optimization exceeds 8%, the material samples provided by the supplier are resampled for destructive testing, and the basic parameter database is updated. When the execution control error exceeds 2mm, the transmission components of the drive mechanism are tested for laser interferometer accuracy, and the ball screws that are worn beyond the tolerance band are replaced. All optimization operations form a closed-loop control. After each adjustment, it must pass the verification test of three complete transportation cycles to ensure that the system stability improvement meets the preset KPI indicators.

[0037] Embodiment 2:

[0038] This embodiment is based on the first embodiment: optimizes the moisture-sensitive material processing flow and strengthens the error tracing mechanism; specifically, the environmental perception module adds a high-precision dew point sensor to monitor the condensation risk on the surface of the packaging box with a resolution of 0.1°C, and triggers the moisture-proof preprocessing instruction when the difference between the dew point temperature and the ambient temperature is detected to be less than 2°C; the LSTM neural network of the data analysis module is upgraded to the spatiotemporal attention model input layer, and the surface deformation gradient field data output layer is added to extend the prediction time window to 4 hours; the moisture-sensitive material determination standard is tightened, and the tightening means that the determination standard for moisture-sensitive materials becomes more stringent, specifically, the material list with a hygroscopic expansion coefficient ≥ 0.8mm / %RH is dynamically updated once an hour.

[0039] When the predicted deformation exceeds 60% of the safety limit, the response mechanism is triggered, and a fourth-level automatic desiccant delivery and a fifth-level cold chain logistics switching scheme are added; a micro hot air gun is installed at the end of the robotic arm of the execution control module, and the contact surface of the packaging box is preheated for 30 seconds before installing the moisture-proof material to eliminate the surface condensation film; adversarial sample generation technology is introduced in the calibration process to simulate extreme humidity mutation scenarios and train the model robustness; the error attribution system adds a supplier batch defect detection function, and automatically freezes the inventory and traces it back to the production batch number when the loss rate of the same batch of materials is abnormally high; the rapid approval process integrates blockchain evidence technology, and all electronic signatures and approval records are stored on the chain after hash encryption; compared with Example 1, this embodiment refines the humidity control strategy and enhances data credibility, so that the accuracy of moisture-proof decision-making is improved by 18%, the abnormal material loss rate is reduced by 27%, and the compliance audit efficiency of the approval process is improved by 45%.

[0040] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent selection system for packaging materials based on multi-dimensional data analysis, characterized in that: It includes environment perception module, data analysis module, decision engine module and execution control module; The environmental perception module collects vibration spectrum, surface strain and temperature and humidity data in real time through the sensor network deployed on the surface of the packaging box, and generates an environmental state matrix after feature extraction. The sensor network includes an automatic calibration unit, wherein the surface strain data is used to construct a deformation gradient field model through a spatial interpolation algorithm to quantify the strain distribution characteristics of the packaging box surface; The data analysis module has a built-in historical transportation case library to store the performance attenuation records of different materials under transportation environments. After receiving the environmental state matrix, it establishes a correlation model between vibration energy distribution and material performance attenuation, combines the inventory turnover rate data in the supply chain database to predict the remaining life of the material, and establishes a data interface with the logistics management system to obtain the transportation task duration and path planning data in real time; The decision engine module dynamically adjusts the priorities of cost, safety and environmental protection goals according to the transportation stage, integrates the material life prediction results with the supplier's capacity data, and generates a multi-level decision tree including automatic execution plans and emergency plans; The execution control module pre-stores the packaging box size tolerance standard, structural deformation safety limit and registered size database; and converts the selected scheme into a material replacement instruction. After driving the actuator to complete the operation, the error between the actual transportation data and the predicted value is collected and fed back to the data analysis module to trigger the dynamic calibration of the model parameters. The model prediction error is divided into environmental perception error, material performance error and execution control error. The supply chain database includes an index table of material moisture-proof properties and a library of safety critical values ​​for hygroscopic expansion coefficients, which records the deformation coefficient, protection level and irreversible deformation threshold of each material under different humidity environments, and marks the humidity sensitivity level. At the same time, dynamic calculations are performed based on the on-time delivery rate and the quality inspection pass rate to construct a supplier credit scoring system, and the credit scoring threshold is set as the μ-2σ value of the normal distribution curve of historical performance data.

2. The intelligent selection system for packaging materials based on multi-dimensional data analysis according to claim 1, wherein: When the vibration spectrum data analysis identifies that the energy value of the frequency band near the material's natural frequency increases by more than the baseline threshold within a preset time window, the environmental perception module sends a resonance risk warning signal to the data analysis module; after the data analysis module responds to the warning, it retrieves the transportation records of the same frequency band energy characteristics in the historical transportation case library, and counts the damage rate data of the corresponding materials. If the damage rate exceeds the safety threshold, it initiates a re-evaluation request for the current recommended plan to the decision engine module, and generates a priority list of alternative materials.

3. The intelligent selection system of packaging materials based on multi-dimensional data analysis according to claim 1, wherein: The data analysis module calculates the matching degree based on the material remaining life prediction curve and the transportation task duration obtained by the logistics management system, and triggers the real-time inventory verification process of the supply chain database when the predicted life is lower than the task duration; If the verification result shows that the inventory is less than 120% of the current transportation demand, the decision engine module automatically lowers the priority level of the material in the recommended plan, and links the supplier database to screen a list of alternative suppliers with sufficient production capacity and logistics timeliness that meets the shortest delivery cycle. The generation of the alternative supplier list must match the material type in the alternative material priority list.

4. An intelligent selection system for packaging materials based on multi-dimensional data analysis according to claim 1, characterized in that: The recommended solutions output by the decision engine module also include material specification parameters and installation coordinate data on the packing box. After receiving the data, the execution control module matches and verifies the material specification parameters with the registered size data of the current packing box; If it is detected that the tolerance between the material size and the installation position of the packing box exceeds the preset safety range, a parameter conflict alarm is returned to the decision engine module, triggering a process to regenerate the recommended solution based on the actual size constraints, and preferentially screening material options with a tolerance compatibility higher than 95%.

5. An intelligent selection system for packaging materials based on multi-dimensional data analysis according to claim 1, characterized in that: After the installation of the new material is completed, the execution control module monitors its initial buffer performance parameters in real time. When the deviation between the measured buffer efficiency and the model prediction value exceeds 5%, a calibration request containing the environmental data timestamp and the material batch code is sent to the data analysis module; the data analysis module extracts the environmental characteristic data of the corresponding time window according to the request, preferentially iterates and updates the coefficients of the performance degradation model of humidity-sensitive materials, and synchronizes the updated model version to the decision engine module.

6. The intelligent selection system for packaging materials based on multi-dimensional data analysis according to claim 1, characterized in that: The credit score in the supply chain database affects the solution generation rules of the decision engine module in real time. When the supplier credit score is lower than the threshold set by the normal distribution curve of the historical performance data, the decision engine module automatically blocks the material options supplied by it; at the same time, the actual material loss rate data recorded by the execution control module is transmitted back to the supply chain database, dynamically adjusting the weight ratio of the on-site performance data in the supplier quality assessment indicators, forming a closed-loop feedback mechanism between the credit score and the actual performance of the material.

7. An intelligent selection system for packaging materials based on multi-dimensional data analysis according to claim 1, characterized in that: When the environmental perception module detects that the rate of change of temperature and humidity exceeds the safety critical value of the material hygroscopic expansion coefficient, the data analysis module immediately starts the calculation of the predicted deformation amount, and evaluates the risk level in combination with the safety limit of the packing box structure deformation; if the predicted deformation amount exceeds 80% of the safety limit, the decision engine module terminates the execution of the current solution and sends an emergency reinforcement instruction to the execution control module. At the same time, based on the material moisture-proof characteristic index table in the supply chain database, a recommended list of alternative materials is generated and a fast approval process is started. The fast approval process includes three steps: automatic compliance verification, electronic signature confirmation by the responsible person, and automatic issuance of the execution instruction.

8. An intelligent selection system for packaging materials based on multi-dimensional data analysis according to claim 1, characterized in that: The data analysis module regularly classifies and attributes the model prediction errors. When the environmental perception error exceeds the preset tolerance three times in a row, the sensor calibration unit of the environmental perception module is triggered to perform accuracy verification; when the cumulative material performance error exceeds the limit, the supplier historical data review process is started and the credit score is recalculated; when the execution control error persists, the control parameter optimization rules of the execution mechanism are dynamically adjusted until the measured operation accuracy reaches the preset standard.

9. The selection method proposed by an intelligent packaging material selection system based on multi-dimensional data analysis according to claim 1, characterized in that: First, the environmental perception module uses the sensor network on the surface of the packing box to collect vibration spectrum, surface strain, and temperature and humidity data in real time to generate an environmental state matrix; the data analysis module combines the historical transportation case library and the supply chain database to predict the remaining life of the material, and docks with the logistics management system to obtain transportation task information; The decision engine module dynamically adjusts the target priority according to the transportation stage, and generates a multi-level decision tree by integrating the material life prediction and supplier production capacity data; The execution control module performs the operation of material replacement according to the selected plan, collects the error between the actual transportation data and the predicted value, and feeds it back to the data analysis module to trigger the dynamic calibration of the model parameters. At the same time, the supply chain database includes a material moisture-proof characteristic index table and a safety critical value library of hygroscopic expansion coefficients, records the deformation coefficients, protection levels and irreversible deformation thresholds of each material in different humidity environments, and constructs a supplier credit scoring system based on the on-time delivery rate and the quality inspection pass rate.