PVC stabilizer performance intelligent evaluation and prediction method and system

Abstract

The invention relates to the technical field of data processing, in particular to a PVC stabilizer performance intelligent evaluation and prediction method and system. The method comprises the steps of collecting test data of a test target, preprocessing the test data, inputting the preprocessed data into a stability evaluation model to train the stability evaluation model, collecting test data of a real-time target, and inputting the test data of the real-time target into the trained stability evaluation model. And obtaining a real-time stability score of the real-time target, finally screening out a target to be detected by combining the real-time stability score of the real-time target, judging whether the target to be detected is the medical PVC, if the target to be detected is the medical PVC, carrying out secondary detection, screening out an unqualified real-time target according to a secondary detection result, and if the target to be detected is the medical PVC, carrying out secondary detection. According to the method, the weights of different test data are adjusted, so that the stability score can reflect different types of stability scores of the PVC stabilizer, secondary detection is performed on the medical PVC, and the high standard of the medical PVC stabilizer is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for intelligently evaluating and predicting the performance of a PVC stabilizer. Background Art

[0002] PVC is the third largest synthetic polymer plastic in the world, with about 40 million tons of PVC produced each year. PVC is a polymer formed by the polymerization of vinyl chloride monomer (VCM) in the presence of initiators such as peroxides and azo compounds or under the action of light or heat according to the free radical polymerization mechanism. Vinyl chloride homopolymers and vinyl chloride copolymers are collectively referred to as vinyl chloride resins.

[0003] A Chinese patent with announcement number CN115761311B discloses a performance detection data analysis method and system for a PVC calcium zinc stabilizer. The method performs image recognition processing on each liquid stabilizer image and the corresponding enhanced liquid stabilizer image included in a stabilizer image set to output an initial performance recognition result corresponding to each liquid stabilizer image and an initial performance recognition result corresponding to each enhanced liquid stabilizer image. The corresponding initial performance recognition results are then fused to output a target performance recognition result corresponding to the liquid PVC calcium zinc stabilizer. However, in the prior art, different components in different types of PVC stabilizers are not combined for calculation, resulting in an incomplete stability judgment of the PVC stabilizer and difficulty in reflecting the stability differences between different types of PVC stabilizers. Summary of the invention

[0004] The purpose of the present invention is to propose a method and system for intelligently evaluating and predicting the performance of a PVC stabilizer in view of the problems existing in the background technology.

[0005] The technical solution of the present invention:

[0006] On the one hand, the present application provides a method for intelligently evaluating and predicting the performance of a PVC stabilizer, comprising:

[0007] Collect test data of the test target, and preprocess the test data to obtain preprocessed data;

[0008] Creating a stability assessment model, inputting the preprocessed data into the stability assessment model to train the stability assessment model, and obtaining a trained stability assessment model;

[0009] Collecting test data of the real-time target, inputting the test data of the real-time target into the trained stability evaluation model, and obtaining the real-time stability score of the real-time target;

[0010] The target to be tested is screened out in combination with the real-time stability score of the real-time target, and it is determined whether the target to be tested is medical PVC. If the target to be tested is medical PVC, a secondary test is performed, and unqualified real-time targets are screened out based on the secondary test results.

[0011] Preferably, collecting test data of the test target and preprocessing the test data to obtain preprocessed data includes:

[0012] Create a test data table;

[0013] Collect test data of the test target and put all collected test data into a test data table; the test data includes target data, static test data and dynamic test data;

[0014] Calculate the average value of each test data separately;

[0015] Determine whether there is missing data in the test data;

[0016] If there is missing data, the missing value is filled with the average value of the corresponding data; the test data after filling the missing value is recorded as preprocessed data; the preprocessed data includes preprocessed static data and preprocessed dynamic data.

[0017] Preferably, a stability assessment model is created, and the preprocessed data is input into the stability assessment model to train the stability assessment model to obtain a trained stability assessment model, including:

[0018] Create a stability assessment model;

[0019] All preprocessed data are divided into training set and test set according to random proportions;

[0020] The training set is input into the stability evaluation model, so that the stability evaluation model continuously learns the relationship between the target data and the target test data, and obtains a trained stability evaluation model; the trained stability evaluation model has the ability to automatically output the stability score of the test target according to the input target data;

[0021] The test set is input into the trained stability assessment model to verify whether the trained stability assessment model is trained.

[0022] Preferably, the training set is input into the stability assessment model so that the stability assessment model continuously learns the relationship between the target data and the target test data to obtain a trained stability assessment model, including:

[0023] Select a test target from the training set and obtain preprocessed data of the test target;

[0024] The stability score of the target is calculated by formula 1;

[0025]

[0026] Among them, Sta is the stability score of the test target, T i is the i-th test data, MFR is the retention rate, K i is the weight corresponding to the i-th test data, ΔY is the stability threshold, and P is the performance decay rate;

[0027] Return to select a target from the training set until all targets in the training set have been selected, and obtain the stability score of each target.

[0028] Preferably, the target to be tested is screened out in combination with the real-time stability score of the real-time target, and it is determined whether the target to be tested is medical PVC. If the target to be tested is medical PVC, a secondary test is performed, and unqualified real-time targets are screened out based on the secondary test results, including:

[0029] Get real-time stability scores for real-time targets;

[0030] All real-time targets are sorted from large to small according to the real-time stability scores, and the last M real-time targets are selected; the selected M real-time targets are recorded as targets to be tested;

[0031] Determine in turn whether each target to be tested is medical PVC;

[0032] If the target to be tested is medical PVC, a secondary test is performed on the target to be tested, and unqualified medical PVC is screened out based on the secondary test result;

[0033] If the target to be tested is not medical PVC, then determine whether the stability score of the target to be tested meets the qualified threshold;

[0034] If the stability score of the target to be tested does not meet the qualified threshold, an early warning is issued.

[0035] Preferably, if the target to be tested is medical PVC, a secondary test is performed on the target to be tested, and unqualified medical PVC is screened out based on the secondary test result, including:

[0036] Select a medical PVC;

[0037] The biocompatibility of the medical PVC is evaluated, and unqualified medical PVC is screened out according to the quality of the biocompatibility evaluation.

[0038] Preferably, the target data includes target components and the content of each component.

[0039] Preferably, the static test data includes static thermal stability, and the dynamic test data includes dynamic thermal stability.

[0040] On the other hand, the present application also provides a PVC stabilizer performance intelligent evaluation and prediction system, including an experimental component, a collection component and a control component, wherein the test target is tested by the experimental component, the test data of the target is collected by the collection component, the PVC stabilizer performance intelligent evaluation and prediction method described in any one of the above is executed by the control component, the test data collected by the collection component is preprocessed by the control component, and the performance of the target is judged in combination with the preprocessed data.

[0041] Preferably, the experimental component includes a static experimental module and a dynamic experimental module, and the static experimental module is used to perform a static test on the test target, and the dynamic experimental module is used to perform a dynamic test on the test target.

[0042] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects:

[0043] The test data of the test target is collected and preprocessed to obtain preprocessed data, and then a stability evaluation model is created. The preprocessed data is input into the stability evaluation model to train the stability evaluation model to obtain a trained stability evaluation model. Then, the test data of the real-time target is collected and input into the trained stability evaluation model to obtain the real-time stability score of the real-time target. Finally, the target to be tested is screened out in combination with the real-time stability score of the real-time target, and it is determined whether the target to be tested is medical PVC. If the target to be tested is medical PVC, a secondary test is performed, and unqualified real-time targets are screened out according to the secondary test results. The present application adjusts the weights of different test data so that the stability score can reflect the stability scores of different types of PVC stabilizers, and performs a secondary test on medical PVC to ensure the high standards of medical PVC stabilizers. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of a process for intelligently evaluating and predicting the performance of a PVC stabilizer proposed by the present invention;

[0045] Figure 2 This is a schematic diagram of the structure of a PVC stabilizer performance intelligent evaluation and prediction system proposed by the present invention;

[0046] Reference numerals: 100, experimental component; 101, static experimental module; 102, dynamic experimental module;

[0047] 200, acquisition component; 300, control component. DETAILED DESCRIPTION

[0048] Embodiment 1, as Figure 1As shown, the present invention proposes a method for intelligently evaluating and predicting the performance of a PVC stabilizer, comprising:

[0049] S100, collecting test data of a test target, and preprocessing the test data to obtain preprocessed data;

[0050] S200, creating a stability assessment model, inputting the preprocessed data into the stability assessment model to train the stability assessment model, and obtaining a trained stability assessment model;

[0051] S300, collecting test data of the real-time target, inputting the test data of the real-time target into the trained stability evaluation model, and obtaining a real-time stability score of the real-time target;

[0052] S400, screening out the target to be tested in combination with the real-time stability score of the real-time target, and determining whether the target to be tested is medical PVC, and if the target to be tested is medical PVC, performing a secondary test, and screening out unqualified real-time targets based on the secondary test results;

[0053] Specifically, there are many types of PVC stabilizers, including lead salts, calcium zinc, organic tin and rare earth stabilizers. The performance of each type of PVC stabilizer is different. For example, calcium zinc PVC stabilizers are more suitable for medical products.

[0054] In the present invention, test data of a test target is collected and preprocessed to obtain preprocessed data, then a stability evaluation model is created, the preprocessed data is input into the stability evaluation model to train the stability evaluation model, and the trained stability evaluation model is obtained, then test data of a real-time target is collected, the test data of the real-time target is input into the trained stability evaluation model, and the real-time stability score of the real-time target is obtained, and finally the target to be tested is screened out in combination with the real-time stability score of the real-time target, and it is determined whether the target to be tested is medical PVC, if the target to be tested is medical PVC, a secondary test is performed, and unqualified real-time targets are screened out according to the secondary test results, and the present application adjusts the weights of different test data so that the stability score can reflect the stability scores of different types of PVC stabilizers, and performs a secondary test on medical PVC to ensure the high standard of medical PVC stabilizers.

[0055] In an optional embodiment, the 100 includes:

[0056] S110, creating a test data table;

[0057] S120, collecting test data of the test target, and putting all collected test data into a test data table; the test data includes target data, static test data and dynamic test data;

[0058] S130, respectively calculating the average value of each test data;

[0059] S140, determining whether there is missing data in the test data;

[0060] Specifically, when searching for missing data, you can use isnull() to implement statistics on missing values;

[0061] S150, if there is missing data, the missing value is filled with the average value of the corresponding data; the test data after the missing value is filled is recorded as preprocessed data; the preprocessed data includes preprocessed static data and preprocessed dynamic data.

[0062] It should be noted that by collecting multiple data of the test target and preprocessing each data to fill in the missing data and ensure the integrity of the data, it is convenient for the subsequent training of the stability assessment model.

[0063] In an optional embodiment, the S200 includes:

[0064] S210, creating a stability assessment model;

[0065] S220, dividing all preprocessed data into a training set and a test set according to a random ratio;

[0066] Specifically, the division ratio of the training set should be greater than that of the test set;

[0067] S230, inputting the training set into the stability evaluation model, so that the stability evaluation model continuously learns the relationship between the target data and the target test data, and obtains a trained stability evaluation model; the trained stability evaluation model has the ability to automatically output the stability score of the test target according to the input target data;

[0068] S240, input the test set into the trained stability assessment model to verify whether the trained stability assessment model is trained.

[0069] It should be noted that by inputting the training samples in the training set into the stability assessment model in sequence, the stability assessment model continuously learns the relationship between the target data and the target's stability score, so that the user can directly input the target data of the test target into the trained stability assessment model to directly obtain the stability score of the test target.

[0070] In an optional embodiment, the S230 includes:

[0071] S231, selecting a test target from the training set and obtaining preprocessed data of the test target;

[0072] S232, calculating the stability score of the target by using Formula 1;

[0073]

[0074] Among them, Sta is the stability score of the test target, T i is the i-th test data, MFR is the retention rate, K i is the weight corresponding to the i-th test data, ΔY is the stability threshold, and P is the performance decay rate;

[0075] Specifically, MFR (melt flow rate) refers to the number of grams of molten resin that flows out through a standard capillary within a certain period of time (usually 10 minutes) under a certain temperature and pressure, and the unit is g / 10min. MFR is an important reference for selecting plastic processing materials and grades, which can help select raw materials that are more suitable for processing technology and improve the reliability and quality of product molding.

[0076] S233, return to step S231, until all targets in the training set are selected, and the stability score of each target is obtained.

[0077] It should be noted that the stability score of the target is calculated by Formula 1. Since the target data includes a variety of test data of PVC stabilizers, and a corresponding weight is set for each test data in Formula 1, the final stability score can reflect the performance of PVC stabilizers in different aspects according to different weights.

[0078] For example, there are PVC stabilizer A and PVC stabilizer B. Now we need to compare the stability difference between PVC stabilizer A and PVC stabilizer B. Since we need to screen out the better PVC stabilizer for transparent hard product processing, we pay more attention to the thermal stability of PVC stabilizer. Therefore, when calculating the stability score of PVC stabilizer, we increase the corresponding weight of the relevant test data of thermal stability. Finally, the stability score of PVC stabilizer A is higher than that of PVC stabilizer B. It can be proved that the thermal stability of PVC stabilizer A is better than the static stability of PVC stabilizer B. Therefore, PVC stabilizer A is more suitable for processing transparent hard products.

[0079] In an optional embodiment, the S400 includes:

[0080] S410, obtaining a real-time stability score of a real-time target;

[0081] S420, sorting all real-time targets from large to small according to the real-time stability scores, and selecting the last M real-time targets; recording the selected M real-time targets as targets to be tested;

[0082] S430, determining in turn whether each target to be tested is medical PVC;

[0083] S440, if the target to be tested is medical PVC, a secondary test is performed on the target to be tested, and unqualified medical PVC is screened out based on the secondary test result;

[0084] S450, if the target to be tested is not medical PVC, determining whether the stability score of the target to be tested meets a qualified threshold;

[0085] S460: If the stability score of the target to be measured does not meet the qualified threshold, a warning is issued.

[0086] It should be noted that since the requirements for medical PVC are higher than those for general PVC, secondary testing of the stabilizer of medical PVC is required to ensure the high standards for medical PVC. When testing the performance of the stabilizer of medical PVC, it is generally judged based on its biocompatibility.

[0087] In an optional embodiment, the S440 includes:

[0088] S441, select a medical PVC;

[0089] S442, performing biocompatibility evaluation on the medical PVC, and screening out unqualified medical PVC according to the quality of the biocompatibility evaluation;

[0090] Specifically, biocompatibility refers to the property of a material not inducing harmful reactions and maintaining functional balance when interacting with a host (such as human tissues and body fluids). This interaction involves the effect of the material on the organism (such as toxicity and inflammatory response) and the feedback of the organism to the material (such as degradation and capsule formation). The interaction between the two parties continues until a dynamic equilibrium is reached or the material is removed.

[0091] It should be noted that when conducting biocompatibility assessment, in addition to testing the composition of PVC stabilizers, chronic toxicity tests can also be performed on some PVC devices to observe the long-term effects of material degradation products on surrounding tissues. Medical PVC stabilizers with poor biocompatibility also need to be classified as unqualified PVC stabilizers and an early warning should be issued.

[0092] In an optional embodiment, the target data includes target components and the content of each component.

[0093] It should be noted that the content of the main components of different types of PVC stabilizers is different. For metal salt heat stabilizers, they mainly include lead salts, calcium-zinc composite stabilizers, barium-cadmium stabilizers, etc., while for organic heat stabilizers, they mainly include stearates, epoxies, phosphites, organic tin, etc.

[0094] In an optional embodiment, the static test data includes static thermal stability, and the dynamic test data includes dynamic thermal stability.

[0095] It should be noted that, specifically, the static thermal stability can be tested by the Congo red method, and the dynamic thermal stability can be tested by the double-roll milling method.

[0096] like Figure 2 As shown, the present application also provides a PVC stabilizer performance intelligent evaluation and prediction system, including an experimental component 100, a collection component 200 and a control component 300, wherein the test target is tested by the experimental component 100, the test data of the target is collected by the collection component 200, the PVC stabilizer performance intelligent evaluation and prediction method described in any one of the embodiments is executed by the control component 300, the test data collected by the collection component 200 is preprocessed by the control component 300, and the performance of the target is judged in combination with the preprocessed data.

[0097] It should be noted that various performance tests on the test target are completed through the experimental component 100, the collection component 200 collects the test data and transmits it to the control component 300, and the control component 300 is provided with a stability evaluation model, and the stability score of the test target is calculated by combining the stability evaluation model with the test data.

[0098] In an optional embodiment, the experiment component 100 includes a static experiment module 101 and a dynamic experiment module 102 . The static experiment module 101 is used to perform a static test on the test target, and the dynamic experiment module 102 is used to perform a dynamic test on the test target.

[0099] It should be noted that the static test module 101 can perform the Congo test method or the static oven test method to test the static performance of the PVC stabilizer, and the dynamic test module 102 can perform the dynamic double-roller mixing method or the torque table mixing method to test the dynamic performance of the PVC stabilizer.

[0100] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto, and various changes can be made within the knowledge scope of technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. A method for intelligent evaluation and prediction of PVC stabilizer performance, characterized in that: include: Collect test data of the test target, and preprocess the test data to obtain preprocessed data; Creating a stability assessment model, inputting the preprocessed data into the stability assessment model to train the stability assessment model, and obtaining a trained stability assessment model; Collecting test data of the real-time target, inputting the test data of the real-time target into the trained stability evaluation model, and obtaining the real-time stability score of the real-time target; The target to be tested is screened out in combination with the real-time stability score of the real-time target, and it is determined whether the target to be tested is medical PVC. If the target to be tested is medical PVC, a secondary test is performed, and unqualified real-time targets are screened out based on the secondary test results.

2. The method for intelligently evaluating and predicting the performance of a PVC stabilizer according to claim 1, characterized in that: Collect test data of the test target and preprocess the test data to obtain preprocessed data, including: Create a test data table; Collect test data of the test target and put all collected test data into a test data table; the test data includes target data, static test data and dynamic test data; Calculate the average value of each test data separately; Determine whether there is missing data in the test data; If there is missing data, the missing value is filled with the average value of the corresponding data; the test data after filling the missing value is recorded as preprocessed data; the preprocessed data includes preprocessed static data and preprocessed dynamic data.

3. A PVC stabilizer performance intelligent evaluation and prediction method according to claim 2, characterized in that: A stability assessment model is created, and the preprocessed data is input into the stability assessment model to train the stability assessment model, and a trained stability assessment model is obtained, including: Create a stability assessment model; All preprocessed data are divided into training set and test set according to random proportions; The training set is input into the stability evaluation model, so that the stability evaluation model continuously learns the relationship between the target data and the target test data, and obtains a trained stability evaluation model; the trained stability evaluation model has the ability to automatically output the stability score of the test target according to the input target data; The test set is input into the trained stability assessment model to verify whether the trained stability assessment model is trained.

4. A PVC stabilizer performance intelligent evaluation and prediction method according to claim 3, characterized in that: The training set is input into the stability assessment model so that the stability assessment model continuously learns the relationship between the target data and the target test data, and a trained stability assessment model is obtained, including: Select a test target from the training set and obtain preprocessed data of the test target; The stability score of the target is calculated by formula 1; Among them, Sta is the stability score of the test target, T i It is i test data, MFR is the retention rate, K i is the weight corresponding to the i-th test data, ΔY is the stability threshold, and P is the performance decay rate; Return to select a target from the training set until all targets in the training set have been selected, and obtain the stability score of each target.

5. A PVC stabilizer performance intelligent evaluation and prediction method according to claim 4, characterized in that: Combined with the real-time stability score of the real-time target, the target to be tested is screened out, and it is determined whether the target to be tested is medical PVC. If the target to be tested is medical PVC, a secondary test is performed, and unqualified real-time targets are screened out based on the secondary test results, including: Get real-time stability scores for real-time targets; All real-time targets are sorted from large to small according to the real-time stability scores, and the last M real-time targets are selected; the selected M real-time targets are recorded as targets to be tested; Determine in turn whether each target to be tested is medical PVC; If the target to be tested is medical PVC, a secondary test is performed on the target to be tested, and unqualified medical PVC is screened out based on the secondary test result; If the target to be tested is not medical PVC, then determine whether the stability score of the target to be tested meets the qualified threshold; If the stability score of the target to be tested does not meet the qualified threshold, an early warning is issued.

6. A PVC stabilizer performance intelligent evaluation and prediction method according to claim 5, characterized in that: If the target to be tested is medical PVC, a secondary test is performed on the target to be tested, and unqualified medical PVC is screened out based on the secondary test results, including: Select a medical PVC; The biocompatibility of the medical PVC is evaluated, and unqualified medical PVC is screened out according to the quality of the biocompatibility evaluation.

7. A PVC stabilizer performance intelligent evaluation and prediction method according to claim 6, characterized in that: The target data includes target components and the content of each component.

8. A PVC stabilizer performance intelligent evaluation and prediction method according to claim 7, characterized in that: The static test data includes static thermal stability, and the dynamic test data includes dynamic thermal stability.

9. A PVC stabilizer performance intelligent evaluation and prediction system according to claim 8, characterized in that: include: An experimental component, through which a test target is tested; A collection component, through which test data of a target is collected; A control component, through which the method for intelligently evaluating and predicting the performance of a PVC stabilizer according to any one of claims 1 to 8 is executed, through which the test data collected by the collection component are preprocessed, and the performance of the target is judged in combination with the preprocessed data.

10. A PVC stabilizer performance intelligent evaluation and prediction system according to claim 9, characterized in that: The experimental component includes a static experimental module and a dynamic experimental module. The static experimental module is used to perform a static test on the test target, and the dynamic experimental module is used to perform a dynamic test on the test target.

Images