A phase change microcapsule design system and method based on artificial intelligence

Abstract

The application provides a phase change microcapsule design system and method based on artificial intelligence, which comprises a machine learning module, a data establishing module, a target input module, a machine screening module and a target output module, the machine learning module is used for collecting raw material composition, synthesis process and performance data of existing phase change microcapsules, the data establishing module is used for collecting types of organic materials and inorganic materials and material preparation processes, the target input module is used for inputting phase change temperature, phase change enthalpy value, coating rate and type of core material of the phase change microcapsule to be designed, the machine screening module is used for screening suitable wall materials and emulsifiers from the data establishing module according to machine learning results and data of the target input module, and the target output module is used for outputting the designed phase change microcapsule.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of phase change materials, and in particular to a phase change microcapsule design system and method based on artificial intelligence. BACKGROUND

[0002] Phase change microcapsules are a special type of microcapsule whose core contains one or more phase change materials. When the ambient temperature changes, these phase change materials absorb or release heat, thereby regulating the temperature around the microcapsule. This property makes phase change microcapsules widely used in many fields, especially in situations where precise temperature control is required. There are many types of wall materials, core materials and emulsifiers used to prepare phase change microcapsules. It is usually necessary to go through a large number of experiments to select the appropriate wall material, core material and emulsifier. For example, the paper "Influence of Different Emulsifiers on the Performance of Stearyl Dodecyl Ester Phase Change Microcapsules" describes the preparation of phase change microcapsules with stearyl dodecyl ester (DS) as the core and melamine-urea-formaldehyde resin (MUF) as the shell by in-situ polymerization using three different emulsifiers: sodium dodecyl benzene sulfonate / polyvinyl alcohol (SDBS / PVA), Tween 20 and octylphenol polyoxyethylene ether (OP-10). The preparation time of different emulsifiers is more than 24 hours, and after preparation, performance characterization is also required, which is time-consuming and costly.

[0003] With the development of modern technologies such as big data, machine learning, artificial intelligence and high-performance computing, it is inevitable trend for future materials science research to study the performance of various materials in a data-driven manner, and to provide reliable support for material property prediction and design through processing and analysis of large-scale data. Against this background, it is very important to provide a phase change microcapsule design system based on artificial intelligence. SUMMARY

[0004] The present application provides a phase change microcapsule design method based on artificial intelligence to solve the problem of complicated design process and the need for a large number of experiments in related technologies.

[0005] To solve the above technical problems, the technical solution adopted by the present application is:

[0006] The application discloses an artificial intelligence-based phase change microcapsule design system, which comprises a machine learning module, a data establishing module, a target input module, a machine screening module and a target output module.

[0007] The data establishing module comprises a material type data unit and a synthesis process data unit.

[0008] The machine screening module comprises a wall material screening unit, an emulsifier screening unit and a synthesis process screening unit.

[0009] The application further discloses a method for using the artificial intelligence-based phase change microcapsule design system.

[0010] S1, sending raw material compositions, synthesis processes and performance data of existing phase change microcapsules to the machine learning module for machine learning;

[0011] S2, collecting types of organic materials and inorganic materials and material preparation processes to establish the material type data unit and the synthesis process data unit;

[0012] S3, providing and inputting phase change temperature, phase change enthalpy, encapsulation rate and core material type of the phase change microcapsule to be designed;

[0013] S4, calculating parameters of wall materials and emulsifiers according to the machine learning result and the phase change temperature, phase change enthalpy, encapsulation rate and core material type of the phase change microcapsule to be designed, and screening wall materials and emulsifiers meeting the requirements from the material type data unit according to the corresponding parameters of the wall materials and the emulsifiers;

[0014] S5, outputting the wall material type, the emulsifier type and the amount of the phase change microcapsule.

[0015] The specific process of the step S1 is as follows.

[0016] Select a certain number of phase change microcapsules from the Internet database, and construct original data sets according to the raw material composition, preparation process and performance parameters of the phase change microcapsules, specifically raw material composition data set, preparation process data set and performance parameter data set;

[0017] The original data set is divided into a training set and a test set according to a set proportion, the data of the training set is used to train the GBRT model, and in the training process, the data of raw material composition and preparation process are used as input objects, and the performance parameter data is used as the target; after the training is completed, the raw material composition data and the preparation process data of the test set are substituted into the machine learning model to obtain the predicted value of the test set; the true value and the predicted value of the test set are compared to evaluate the prediction performance of the machine learning model;

[0018] On the basis of the machine learning model, a data-driven method SISSO is used to construct a mathematical model of the performance parameters of the phase change microcapsules, specifically:

[0019] ;

[0020] ;

[0021] ;

[0022] In the formula, F(CR) represents the coating rate of the phase change microcapsule, X b1 represents the density of the wall material, the unit is g / cm 3 , X r1 represents the molecular weight of the emulsifier, the unit is g / mol, X r2 represents the density of the emulsifier, the unit is g / cm 3 , X c represents the dosage ratio of the core material and the wall material, X t represents the reaction temperature, the unit is ℃; F(PTT) represents the phase change temperature of the phase change microcapsule, the unit is ℃, X x1 represents the density of the core material, the unit is g / cm 3 ; X x2 represents the melting point of the core material, the unit is ℃, X b2 represents the specific heat capacity of the wall material, the unit is J / (g·℃), X x3 represents the specific heat capacity of the core material, the unit is J / (g·℃); X r3 represents the mass percentage of the emulsifier; X r4 represents the HLB value of the emulsifier; X b3 represents the molecular weight of the wall material, the unit is g / mol; F(LHPC) represents the phase change enthalpy value of the phase change microcapsule, the unit is J / g.

[0023] The specific process of step S4 is as follows:

[0024] According to the data of the coating rate, phase change temperature, phase change enthalpy value of the phase change microcapsule to be designed and the type of the core material, combined with the mathematical model constructed in S1, iterative calculation is performed, and the results of two consecutive iterations meet: AF(CR) ≤±0.005, AF(PTT) ≤±0.5, AF(LHPC) ≤±1, that is, stop, and the density of the wall material, the specific heat capacity of the wall material, the molecular weight of the wall material, the molecular weight of the emulsifier, the density of the emulsifier, the mass percentage of the emulsifier, the HLB value of the emulsifier, the dosage ratio of the core material and the wall material and the reaction temperature are obtained, the specific type of the wall material is determined according to the density of the wall material, the specific heat capacity of the wall material and the molecular weight of the wall material, and the specific type of the emulsifier is determined according to the density of the emulsifier, the HLB value of the emulsifier and the molecular weight of the emulsifier; the dosage of the wall material and the core material in the preparation process is determined according to the dosage ratio of the core material and the wall material; the dosage of the emulsifier is determined according to the mass percentage of the emulsifier; and the temperature required in the reaction process is determined according to the reaction temperature.

[0025] In the above step S1, 1000 phase change microcapsules are selected from an Internet database; and the original data set is divided into a training set and a test set according to a ratio of 80:20.

[0026] The application provides a phase change microcapsule design system and method based on artificial intelligence, which has the following beneficial effects:

[0027] In the traditional method, the process of designing a phase change microcapsule meeting the performance requirements according to the requirements is complicated, a large amount of literature data needs to be manually queried, experimental verification needs to be performed, and a long time is consumed; the design module provided in the application adopts a machine learning method to design a phase change microcapsule, which effectively saves the labor cost and shortens the research and development time, and the system provided in the application can complete the design of a phase change microcapsule material meeting the composite performance requirements in 10-20 seconds. BRIEF DESCRIPTION OF DRAWINGS

[0028] The application will be further described below in combination with the drawings and examples:

[0029] Figure 1 A flowchart of the method of the phase change microcapsule design system based on artificial intelligence provided in the application is shown in the figure;

[0030] Figure 2 A microscope diagram of the phase change microcapsule designed by using the phase change microcapsule design system based on artificial intelligence provided in the application is shown in the figure;

[0031] Figure 3 A phase change enthalpy value diagram of the phase change microcapsule designed by using the phase change microcapsule design system based on artificial intelligence provided in the application is shown in the figure. DETAILED DESCRIPTION

[0032] In order to make the objects, technical solutions and advantages of the present application clearer, the following will combine the drawings provided by the present application to make a systematic and complete description of the specific technical solutions of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0033] An artificial intelligence-based phase change microcapsule design system comprises a machine learning module, a data establishing module, a target input module, a machine screening module and a target output module. The machine learning module is used to collect existing phase change microcapsule raw material compositions, synthesis processes and performance data. The data establishing module is used to collect organic material, inorganic material types and material preparation processes. The target input module is used to input phase change temperature, phase change enthalpy value, coating rate and core material type of the phase change microcapsule to be designed. The machine screening module is used to screen suitable wall materials and emulsifiers from the data establishing module according to the machine learning results and the data of the target input module. The target output module is used to output the designed phase change microcapsule.

[0034] The above-mentioned data establishing module comprises a material type data unit and a synthesis process data unit. The material type data unit comprises existing inorganic material and organic material types and corresponding parameter data. The synthesis process data unit comprises existing material preparation processes and preparation conditions.

[0035] The above-mentioned machine screening module comprises a wall material screening unit, an emulsifier screening unit and a synthesis process screening unit. The wall material screening unit screens the required wall material from the material type data unit. The emulsifier screening unit screens the required emulsifier from the material type data unit. The synthesis process screening unit screens the required synthesis process from the synthesis process data unit according to the screened wall material and emulsifier types.

[0036] Reference Figure 1 A method using the above-mentioned artificial intelligence-based phase change microcapsule design system comprises the following steps:

[0037] S1, sending existing phase change microcapsule raw material compositions, synthesis processes and performance data to the machine learning module for machine learning;

[0038] S2, collecting organic material, inorganic material types and material preparation processes to establish a material type data unit and a synthesis process data unit;

[0039] S3, providing and inputting phase change temperature, phase change enthalpy value, coating rate and core material type of the phase change microcapsule to be designed;

[0040] S4, according to the machine learning result and the phase change temperature, the phase change enthalpy value, the coating rate and the type of the core material of the phase change microcapsule to be designed, the parameters of the wall material and the emulsifier are calculated, and the wall material and the emulsifier meeting the requirements are screened out from the material type data unit according to the corresponding parameters of the wall material and the emulsifier;

[0041] S5, output the wall material, the emulsifier type and the amount of the phase change microcapsule.

[0042] The specific process of the above step S1 is:

[0043] 1000 kinds of phase change microcapsules are selected from the Internet database, and original data sets are constructed according to the raw material composition, preparation process and performance parameters of these phase change microcapsules, specifically raw material composition data set, preparation process data set and performance parameter data set;

[0044] The original data set is divided into a training set and a test set according to a ratio of 80:20, the data of the training set is used to train the GBRT model, and in the training process, the data of the raw material composition and the preparation process are taken as the input object, and the performance parameter data are taken as the target; after the training is completed, the raw material composition data and the preparation process data of the test set are substituted into the machine learning model to obtain the predicted value of the test set; the true value and the predicted value of the test set are compared to evaluate the prediction performance of the machine learning model, the average absolute error MAE of the machine learning model is 0.058eV / atom, the determination coefficient R 2 =0.87, which indicates that the machine learning model has high accuracy;

[0045] On the basis of the machine learning model, a data-driven method SISSO is used to construct a mathematical model of the performance parameters of the phase change microcapsule, specifically:

[0046] ;

[0047] ;

[0048] ;

[0049] In the formula, F(CR) represents the coating rate of the phase change microcapsule, X b1 represents the density of the wall material, the unit is g / cm 3 , X r1 represents the molecular weight of the emulsifier, the unit is g / mol, X r2 represents the density of the emulsifier, the unit is g / cm 3 , X c represents the amount ratio of the core material and the wall material, X t represents the reaction temperature, the unit is ℃; F(PTT) represents the phase change temperature of the phase change microcapsule, the unit is ℃, X x1Density of the core material, unit: g / cm 3 Density of the core material, unit: g / cm x2 Density of the core material, unit: g / cm b2 Specific heat capacity of the wall material, unit: J / (g·℃) x3 Specific heat capacity of the core material, unit: J / (g·℃) r3 Mass percentage of the emulsifier r4 HLB value of the emulsifier b3 Molecular weight of the wall material, unit: g / mol; F(LHPC) represents the phase transition enthalpy value of the phase change microcapsule, unit: J / g.

[0050] The SISSO can combine machine learning models to find a few important features from various features that represent the performance parameters of the phase change microcapsule. In the present application, the most important features representing the coating rate of the phase change microcapsule are: the density of the wall material, the molecular weight of the emulsifier, the density of the emulsifier, the mass percentage of the emulsifier, the molecular weight of the wall material, and the dosage ratio of the core material and the wall material. The most important features representing the phase transition temperature of the phase change microcapsule are: the density of the core material, the melting point of the core material, the density of the emulsifier, the density of the wall material, the reaction temperature, the dosage ratio of the core material and the wall material, and the mass percentage of the emulsifier. The most important features representing the phase transition enthalpy value of the phase change microcapsule are: the dosage ratio of the core material and the wall material, the specific heat capacity of the wall material, the mass percentage of the emulsifier, the reaction temperature, the HLB value of the emulsifier, the specific heat capacity of the core material, and the melting point of the core material.

[0051] The specific process of the above step S4 is as follows:

[0052] According to the data of the coating rate, the phase transition temperature, and the phase transition enthalpy value of the phase change microcapsule to be designed and the type of the core material, the mathematical model constructed in S1 is combined to perform iterative calculation. The results of two consecutive iterations meet: ΔF(CR) ≤±0.005, ΔF(PTT) ≤±0.5, and ΔF(LHPC) ≤±1, and then the calculation is stopped. The density of the wall material, the specific heat capacity of the wall material, the molecular weight of the wall material, the molecular weight of the emulsifier, the density of the emulsifier, the mass percentage of the emulsifier, the HLB value of the emulsifier, the dosage ratio of the core material and the wall material, and the reaction temperature are obtained. The specific type of the wall material is determined according to the density of the wall material, the specific heat capacity of the wall material, and the molecular weight of the wall material. The specific type of the emulsifier is determined according to the density of the emulsifier, the HLB value of the emulsifier, and the molecular weight of the emulsifier. The dosage of the wall material and the core material in the preparation process is determined according to the dosage ratio of the core material and the wall material. The dosage of the emulsifier is determined according to the mass percentage of the emulsifier. The temperature required in the reaction process is determined according to the reaction temperature.

[0053] Example 1:

[0054] (1) In the Internet database according to the keyword "phase change microcapsule" screening retrieval 1000 groups of phase change microcapsule raw material composition, synthesis process and performance data, and the retrieval results are exported in table form and sent to the machine learning module;

[0055] (2) Collect all existing inorganic materials and organic material species and corresponding parameter data, material preparation process and preparation conditions, import into the data establishment module;

[0056] (3) In the target input module, the phase change temperature of the phase change microcapsule to be designed is set to 30-40℃, the phase change enthalpy value is set to 180-220J / g, the coating rate is set to 85%-95%, and the core material type is set to polyethylene glycol 1000;

[0057] (4) The machine screening module screens out suitable wall materials and emulsifiers from the data establishment module according to the data of the target input module;

[0058] (5) The target output module outputs the wall material, emulsifier type and dosage of the designed phase change microcapsule.

[0059] Using the system provided by the application, after inputting the phase change temperature, phase change enthalpy value, coating rate and core material, after 15s, the output result of the target output module is: the wall material is gelatin with a molecular weight of 60,000, the emulsifier is sodium dodecyl sulfate, the reaction temperature X t =60℃, the dosage ratio of core material and wall material X c =0.4, the mass percentage of emulsifier X r3 =0.03, the coating rate of phase change microcapsule F(CR)=90.2%, the phase change temperature of phase change microcapsule F(PTT)=39.27℃, and the phase change enthalpy value of phase change microcapsule F(LHPC)=210.16J / g.

[0060] Experimental verification:

[0061] According to the wall material type, emulsifier type, dosage ratio of core material and wall material, and mass percentage of emulsifier provided in Example 1, phase change microcapsules are prepared by emulsion polymerization, and the performance of the prepared phase change microcapsules is tested. The coating rate of the prepared phase change microcapsules is 89.5%, the microscope display graph of the phase change microcapsules is shown in Figure 2 , and the phase change enthalpy graph is shown in Figure 3 From Figure 3 , it can be seen that the phase change temperature of the phase change microcapsules prepared according to the raw materials provided by the design system is 39.69℃, and the phase change enthalpy value is 209.29J / g, which is highly consistent with the phase change temperature and phase change enthalpy value output by the design system, indicating that the design system provided by the application can accurately complete the design of the phase change enthalpy value of the phase change microcapsules.

Claims

1. A phase change microcapsule design system based on artificial intelligence, characterized in that, It includes a machine learning module, a data building module, a target input module, a machine screening module, and a target output module. The machine learning module is used to collect data on the raw material composition, synthesis process, and performance of existing phase change microcapsules. The data building module is used to collect the types of organic and inorganic materials and their preparation processes. The target input module is used to input the phase change temperature, phase change enthalpy, encapsulation rate, and core material type of the phase change microcapsule to be designed. The machine screening module is used to select suitable wall materials and emulsifiers from the data building module based on machine learning results and data from the target input module; the target output module is used to output the designed phase change microcapsules. The data establishment module includes a material type data unit and a synthesis process data unit. The material type data unit includes existing inorganic and organic material types and corresponding parameter data, while the synthesis process data unit includes existing material preparation processes and preparation conditions.

2. The phase change microcapsule design system based on artificial intelligence according to claim 1, characterized in that, The machine screening module includes a wall material screening unit, an emulsifier screening unit, and a synthesis process screening unit. The wall material screening unit selects wall materials that meet the requirements from the material type data unit, the emulsifier screening unit selects emulsifiers that meet the requirements from the material type data unit, and the synthesis process screening unit selects synthesis processes that meet the requirements from the synthesis process data unit based on the selected wall materials and emulsifier types.

3. The method using the artificial intelligence-based phase change microcapsule design system as described in claim 2, characterized in that, Includes the following steps: S1. Send the existing raw material composition, synthesis process and performance data of phase change microcapsules to the machine learning module for machine learning; S2. Collect the types of organic and inorganic materials and their preparation processes, and establish data units for material types and synthesis processes. S3. Provide and input the phase transition temperature, phase transition enthalpy, encapsulation rate, and core material type of the phase change microcapsule to be designed; S4. Based on the machine learning results and the phase change temperature, phase change enthalpy, encapsulation rate and core material type of the phase change microcapsule to be designed, calculate the parameters of the wall material and emulsifier. Based on the parameters corresponding to the wall material and emulsifier, select the wall material and emulsifier that meet the requirements from the material type data unit. S5. Output the wall material, type and amount of emulsifier for the phase change microcapsules.

4. The method for designing a phase change microcapsule system based on artificial intelligence according to claim 3, characterized in that, The specific process of step S1 is as follows: A set number of phase change microcapsules of a certain type were selected from the Internet database. Based on the raw material composition, preparation process and performance parameters of these phase change microcapsules, original datasets were constructed respectively, specifically raw material composition dataset, preparation process dataset and performance parameter dataset. The original dataset is divided into training and testing sets according to a set ratio. The GBRT model is trained using the data in the training set. During the training process, the data on raw material composition and preparation process are used as input objects, and the performance parameter data are used as targets. After training, the raw material composition data and preparation process data of the test set are substituted into the machine learning model to obtain the predicted values ​​of the test set; the actual values ​​of the test set are compared with the predicted values ​​to evaluate the predictive performance of the machine learning model. Based on the machine learning model, a mathematical model of the performance parameters of phase change microcapsules is constructed using the data-driven method SISSO, specifically as follows: In the formula, F(CR) represents the encapsulation rate of the phase change microcapsules, and X... b1 The density of the wall material is expressed in g / cm³. 3 ,X r1 X represents the molecular weight of the emulsifier, expressed in g / mol. r2 The density of the emulsifier is expressed in g / cm³. 3 ,X c X represents the ratio of core material to wall material usage. t The reaction temperature is expressed in °C; F(PTT) represents the phase transition temperature of the phase change microcapsules, also expressed in °C; X x1 The density of the core material is expressed in g / cm³. 3 ;X x2 Indicates the melting point of the core material, in °C, X b2 X represents the specific heat capacity of the wall material, expressed in J / (g·℃). x3 This indicates the specific heat capacity of the core material, expressed in J / (g·℃); X r3 Indicates the mass percentage of emulsifier; X r4 Indicates the HLB value of the emulsifier; X b3 The value represents the molecular weight of the wall material, expressed in g / mol; F(LHPC) represents the phase transition enthalpy of the phase change microcapsules, expressed in J / g.

5. The method for designing a phase change microcapsule system based on artificial intelligence according to claim 4, characterized in that, The specific process of step S4 is as follows: Based on the encapsulation rate, phase transition temperature, phase transition enthalpy, and core material type of the phase change microcapsules to be designed, iterative calculations are performed using the mathematical model constructed in S1. The calculations stop when the results of two consecutive iterations satisfy: ΔF(CR)≤±0.005, ΔF(PTT)≤±0.5, ΔF(LHPC)≤±1. This yields the density, specific heat capacity, molecular weight, molecular weight, density, mass percentage, HLB value, core-to-wall material ratio, and reaction temperature of the wall material. The specific wall material type is determined based on its density, specific heat capacity, and molecular weight; the specific emulsifier type is determined based on its density, HLB value, and molecular weight; the amount of wall material and core material used in the preparation process is determined based on the core-to-wall material ratio; and the amount of emulsifier is determined based on its mass percentage. The temperature required for the reaction process is determined based on the reaction temperature.

6. The method for designing a phase change microcapsule system based on artificial intelligence according to claim 5, characterized in that, In step S1, 1000 phase transition microcapsules are selected from an internet database; the original dataset is divided into a training set and a test set in an 80:20 ratio.

Images