Quantitative structure-property relationship natural ester molecular structure modification and performance improvement method

Abstract

The application provides a natural ester molecular structure modification and performance improvement method of quantitative structure-property relationship, and the method comprises the following steps: (1) exploring the correlation mechanism between the molecular structure of a natural ester insulating liquid and electrical properties; (2) building a simulation platform for predicting the low-temperature kinematic viscosity and pour point of the natural ester insulating liquid based on molecular dynamics, and based on the simulation prediction results, a prediction model of the low-temperature kinematic viscosity and pour point of the natural ester insulating liquid is built, the quantitative fitting characteristic relationship of the insulating liquid is obtained, and thus the influence mechanism of the molecular configuration and components of the natural ester insulating liquid on the cold flow characteristics is studied; and (3) performing natural ester insulating liquid intrinsic molecular structure modification based on the quantitative structure-property relationship. The application can reveal the molecular relationship between the intrinsic molecular structure of the natural ester insulating liquid and the electrical and cold flow characteristics, and based on this, the structure of the natural ester molecule is modified and the performance is improved. This has important significance for promoting and popularizing the engineering application of the natural ester insulating liquid.

Description

Technical Field

[0001] This invention relates to the fields of quantum chemistry, molecular simulation and artificial intelligence, and specifically to a method for modifying the molecular structure and improving the performance of natural esters by quantitatively analyzing the structure-property relationship. Background Technology

[0002] Transformer insulating fluid is a crucial component of oil-immersed transformers, and its performance significantly impacts the safe and reliable supply of electricity and the transformer's lifespan. Power transformers typically use mineral oil as the insulating fluid. However, mineral oil's poor biodegradability and low flash point limit its application. Due to environmental protection and fire safety requirements, natural ester insulating fluid offers unique advantages over traditional mineral insulating oil. These advantages include: renewability, high biodegradability, and environmental friendliness; a high flash point and ignition point providing excellent fire resistance; a higher dielectric constant improving the uniformity of the electric field distribution in oil-paper insulation; and a saturated water content approximately 20 times that of mineral insulating oil, resulting in less moisture impact and higher tolerance to environmental moisture. Therefore, natural ester insulating fluid is considered a good alternative to mineral insulating oil.

[0003] However, the research, development, and widespread use of natural ester insulating fluids currently face key shortcomings. In terms of electrical performance, the breakdown voltage of natural ester insulating fluids drops sharply over long gaps caused by lightning strikes; its accelerating voltage is typically lower than that of mineral oil but equal to its breakdown voltage, and it exhibits faster propagation speed and greater branching density within the ester fluid. Compared to mineral oil, natural ester insulating fluids have weaker electrical withstand capabilities, especially under rapid transient pulse stress. This indicates that natural ester insulating fluids have lower resistance to lightning overvoltages and are more prone to rapid and high-energy discharges. Regarding cold flow characteristics, the high viscosity and high pour point of natural ester insulating fluids reduce their fluidity at low temperatures, making them more susceptible to damage to internal mechanical rotating parts of transformers. Furthermore, higher viscosity reduces the circulation of the insulating oil, resulting in poorer heat transfer and potentially leading to excessively high temperature rises and hot spot temperatures in transformers during operation.

[0004] In summary, the inherent electrical performance defects and poor cold flow characteristics of natural ester insulating fluids make them unsuitable for use in high-capacity, high-voltage transformers and in cold climates. Therefore, exploring modification methods to improve the electrical and cold flow properties of insulating fluids from the perspective of the quantitative structure and properties of natural ester molecules has significant engineering implications. Summary of the Invention

[0005] The purpose of this invention is to provide a method for modifying the molecular structure and improving the performance of natural esters by quantitatively elucidating their structure-property relationships. This method reveals the molecular mechanism by which the intrinsic molecular structure of natural ester insulating liquids affects their electrical and cold flow properties, and based on this, the structure of natural ester molecules can be modified and their performance improved. This provides important fundamental theoretical support and practical application value for promoting and expanding the engineering applications of natural ester insulating liquids.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for modifying the structure and improving the performance of natural ester molecules based on quantitative structure-property relationships, wherein the natural ester molecules are triglyceride molecules, mainly including triglyceride trilinolenic acid, triglyceride trilinoleic acid, triglyceride trioleate, triglyceride tristearate, and triglyceride tripalmitate.

[0008] The method for modifying and improving the molecular structure of natural esters based on quantitative structure-property relationships aims to improve the performance of natural ester insulating liquids.

[0009] The method for modifying and improving the molecular structure and properties of natural esters based on quantitative structure-property relationships mainly includes improved electrical properties and cold flow properties.

[0010] A method for modifying the structure and improving the performance of natural ester molecules based on quantitative structure-property relationships, characterized in that the natural ester molecule is a triglyceride molecule, which mainly includes triglyceride trilinolenic acid ester, triglyceride trilinoleic acid ester, triglyceride tristearate ester, and triglyceride tripalmitate or others.

[0011] The purpose of the molecular structure modification is to improve the performance of the natural ester insulating liquid. The improved performance mainly includes electrical properties and cold flow properties.

[0012] Preferably, the process includes the following steps: a) establishing the correlation between the molecular structure and electrical properties of natural ester insulating liquid; b) establishing the molecular dynamics prediction of the cold flow characteristics of natural ester and the influence of molecular configuration and components; c) performing intrinsic molecular structure modification of natural ester insulating liquid based on quantitative structure-property relationships.

[0013] More preferably, establishing the correlation between the molecular structure and electrical properties of natural ester insulating liquid specifically includes the following steps:

[0014] a.1) The molecular structure of natural ester insulating liquid molecules under the implicit solvent model was optimized using the quantum chemistry software Gaussian at the B3LYP-D3 / 6-311G(d,p) computational level, and frequency vibration calculations were performed to obtain a high-precision molecular structure without imaginary frequencies.

[0015] a.2) Calculate the electronic structure energy at the high-precision calculation level of M06-2X / def2-TZVPP, and calculate the electronic excited state at the high-precision level of M06-2X / aug-cc-pVTZ. By analyzing the above calculation results, obtain the electronic structure property parameters and molecular structure geometric parameters that are strongly correlated with the molecular dielectric properties.

[0016] a.3) Calculate the parameter changes after applying electric fields of different strengths, and analyze and compare the reasons and rules for the different dielectric properties of natural ester molecules due to different structural characteristics based on the parameter change law;

[0017] a.4) Calculate the global exponent, real space function, atomic exponent, or other quantities reflecting the reactivity of chemical substances and reaction sites in the concept density functional theory using the quantum chemical wavefunction analysis program Multiwfn;

[0018] The correlation between the molecular structure and electrical properties of natural ester insulating liquid was established.

[0019] More preferably, in step a.2), the electronic structure property parameters mainly include electrical and photoelectric performance indicators such as ionization energy, electron affinity, electron density, dipole moment, average local ionization energy, polarizability, transition excitation energy, oscillator strength, band gap, nucleophilicity, and electrophilic hyperdelocalization; the molecular structure geometric parameters mainly include bond length, bond angle, dihedral angle, atomic type and number, and chemical bond type and number.

[0020] More preferably, establishing the molecular dynamics prediction of the cold flow properties of natural esters and the influence relationship between molecular configuration and components specifically includes the following steps:

[0021] b.1) A molecular model of the actual natural ester insulating liquid was established using the molecular dynamics initial structure building tool Packmol, and the molecular dynamics simulation software GROMACS was selected to study the simulation prediction of the low-temperature kinematic viscosity and pour point of the natural ester insulating liquid.

[0022] b.2) Establish a quantitative fitting relationship between the cold flow characteristics of natural ester insulating liquid and temperature by combining the free volume theory;

[0023] b.3) We will use IGM and RDG-based methods to investigate the weak interactions between and within different natural ester molecules and analyze the influence of changes in the molecular structure of natural esters on the weak interactions.

[0024] b.4) Based on the above-mentioned influence laws, a molecular dynamics prediction model for the low-temperature kinematic viscosity of natural ester molecules with different proportions was constructed and compared with the cold flow characteristic parameters of actual natural ester insulating liquid. The study obtained the molecular component mixing rules that can reduce the low-temperature kinematic viscosity and pour point of the natural ester insulating liquid molecular system.

[0025] More preferably, in step b.1), the molecular dynamics calculation settings mainly include using the OPLS-AA force field to describe the potential energy of the natural ester molecular system and the forces between individual atoms; using the Leap-frog method for dynamic simulation and setting periodic boundary conditions in all three directions of the system; using the Particle-Mesh Ewald (PME) method to handle electrostatic interactions, with the short-range electrostatic interaction distance being 1–2 nm. Van der Waals interactions are calculated using a truncation method, with a cutoff value also being 1–2 nm; and using EMD (such as equilibrium cross-flow autocorrelation function) and NEMD (such as non-equilibrium periodic perturbation method) methods to calculate the cold flow characteristic parameters of the natural ester insulating liquid.

[0026] More preferably, in step b.3), the weak interactions between different natural ester molecules mainly include hydrogen bonds, van der Waals interactions, electrostatic interactions, or others; the weak interactions within molecules mainly include hydrogen bonds, dihydrogen bonds, van der Waals interactions, or others; and the molecular structural characteristics include characteristic functional groups, molecular chain length, number of molecular branches, length of molecular branches, molecular saturation, or others.

[0027] Preferably, the intrinsic molecular structure modification of natural ester insulating liquid based on quantitative structure-property relationships specifically includes the following steps:

[0028] c.1) Utilize the integrated network-based molecular descriptor and fingerprint computing platform alvaDesc to calculate the molecular topological descriptors, quantum chemical property descriptors, and physicochemical property descriptors in the quantitative structure-property relationship of natural ester molecules;

[0029] c.2) Construct a massive database of QSPR natural ester molecule descriptors based on SMILES molecular fingerprint characterization and precise 3D molecular structure characterization;

[0030] c.3) The quantitative structure-property relationship between natural ester molecule descriptors and predicted properties is studied by computer-aided design of high-throughput virtual screening. Combined with statistical analysis, evolutionary algorithms or other methods, the preprocessing and efficient optimization of natural ester molecule descriptors are achieved.

[0031] c.4) Use tools such as molecular structure generators to add atoms or modify fragments to change the initial conformation and descriptor probability distribution of natural ester molecules. Combine this with molecular structure weak site analysis to quantify the correlation between molecular structure feature parameters and target properties. Use predictive models to redistribute and screen descriptors to achieve the desired targeted property molecular modification design and obtain the optimal modified natural ester molecule structure.

[0032] More preferably, in step c.1), the molecular topological descriptor mainly includes molecular topological information such as atomic distance, adjacency relationship, molecular shape, or molecular chemical information such as atomic charge, orbital information, or others; the quantum chemical property descriptor is mainly based on density functional theory calculation; and the physicochemical property descriptor is based on experimental data and collected from literature.

[0033] More preferably, in step c.3), the quantitative structure-property relationship construction method includes neural networks or others; the statistical analysis method mainly includes principal component analysis (PCA), partial least squares (PLS) or others; the evolutionary algorithm mainly includes genetic algorithm (GA), differential evolution algorithm (DE) or others.

[0034] The beneficial effects of this invention are as follows:

[0035] This invention provides a method for modifying and improving the molecular structure of natural esters by quantitatively analyzing their structure-property relationships. Starting with the intrinsic molecular structure of natural ester insulating liquids, the intrinsic correlation mechanism between the molecular structure and electrical properties of natural ester insulating liquids can be revealed, as well as the influence of molecular configuration and components on the cold flow characteristics of natural ester insulating liquids. Furthermore, by combining analysis of weak sites in the molecular structure, the inherent defects in the electrical and cold flow characteristics of natural ester insulating liquids can be improved and resolved, enabling the scientific determination of the desired natural ester molecular structure modification design. This invention utilizes quantum chemistry, molecular simulation, and artificial intelligence technologies, overcoming the shortcomings of traditional methods that rely on extensive macroscopic experimental analysis for natural ester modification, such as complexity, high cost, and long cycle time. It is expected to overcome the technical bottleneck of applying environmentally friendly insulating liquids in a wider range of scenarios, including high-capacity, high-voltage transformers and cold climate environments, and has significant fundamental theoretical exploration needs and practical engineering implications. Attached Figure Description

[0036] Figure 1 The optimized high-precision molecular structure of triglyceride molecules is shown in (a)-(d), which are triglyceride tristearate, triglyceride trioleate, triglyceride trilinoleate and triglyceride trilinolenic acid ester respectively.

[0037] Figure 2 Isosurface plots of electron density differences before and after applying electric fields of different intensities to triglycerides.

[0038] Figure 3 This is a molecular model of a natural ester insulating liquid.

[0039] Figure 4 Surface analysis diagram of the van der Waals potential of glyceryl tristearate.

[0040] Figure 5 A weighted ranking graph of natural ester molecule descriptors.

[0041] Figure 6Flowchart for constructing quantitative structure-property relationships.

[0042] Figure 7 Example structure of a modified natural ester molecule. Detailed Implementation

[0043] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments, but the scope of protection of the present invention is not limited thereto.

[0044] Example 1

[0045] A quantitative structure-property relationship-based method for modifying the molecular structure and improving the performance of natural esters includes the following steps: a) exploring the correlation mechanism between the molecular structure and electrical properties of natural ester insulating liquids; b) studying the molecular dynamics prediction of the cold flow characteristics of natural esters and the influence mechanism of molecular configuration and components; c) conducting research on the intrinsic molecular structure modification of natural ester insulating liquids based on quantitative structure-property relationships.

[0046] a) Investigating the correlation mechanism between the molecular structure and electrical properties of natural ester insulating liquids specifically includes the following steps:

[0047] 1) The molecular structures of glycerol tristearate, glycerol trioleate, glycerol trilinoleate, and glycerol trilinolenic acid esters under an implicit solvent model were optimized using the quantum chemistry software Gaussian at the B3LYP-D3 / 6-311G(d,p) computational level. Frequency vibration calculations were performed to obtain high-precision molecular structures without imaginary frequencies, such as... Figure 1 As shown;

[0048] 2) Electronic structure energy was calculated using the high-precision M06-2X / def2-TZVPP method, and electronic excited states were calculated using the high-precision M06-2X / aug-cc-pVTZ method. By analyzing the above calculation results, electronic structure property parameters that are strongly correlated with molecular dielectric properties were obtained, such as ionization energy, electron affinity, electron density, dipole moment, average local ionization energy, polarizability, transition excitation energy, oscillator strength, band gap, nucleophilicity and electrophilic hyperdelocality, as well as molecular structure geometric parameters such as bond length, bond angle, dihedral angle, atom type and number, and chemical bond type and number. Some calculation results are shown in Table 1.

[0049] Table 1. Partial electronic structure and geometric parameters of triglyceride molecules

[0050]

[0051] 3) Calculate the parameter changes after applying different electric fields. Based on the parameter change patterns, analyze and compare the reasons and patterns for the different dielectric properties of natural ester molecules due to their different structural characteristics. For example... Figure 2 The figure shows the isosurface plots of electron density differences in trioleate before and after applying electric fields of different intensities. The green and blue areas can be considered as electron and hole regions, respectively. It is evident from the figure that as the electric field strength increases, the molecular polarization becomes more pronounced, eventually leading to the formation of an internal electric field within each molecule, opposite in direction to the external electric field. Furthermore, with increasing electric field strength, the space charge of trioleate exhibits a significant branching phenomenon along the molecular structure. This may explain why natural ester insulating liquids are observed to have more dendritic branching compared to mineral insulating oils during negative flow discharge.

[0052] 4) The global index, real space function, and atomic index reflecting the reactivity of chemical substances and reaction sites in the concept density functional theory were calculated using the quantum chemical wave function analysis program Multiwfn. For example, some real space function calculation results are shown in Table 2.

[0053] Table 2. Calculation results of partial real space functions of triglyceride molecules.

[0054]

[0055] b) The study of the molecular dynamics prediction of the cold flow properties of natural esters and the influence mechanism of molecular configuration and components specifically includes the following steps:

[0056] 1) A molecular model of a real natural ester insulating liquid was established using the molecular dynamics initial structure building tool Packmol, such as... Figure 3 As shown in Table 3, the molecular dynamics simulation software GROMACS was used to study the simulation prediction of the low-temperature kinematic viscosity and pour point of the natural ester insulating liquid. The OPLS-AA force field was used to describe the potential energy of the natural ester molecular system and the forces between individual atoms. The Leap-frog method was used for dynamic simulation, and periodic boundary conditions were set in all three directions of the system. The Particle-Mesh Ewald (PME) method was used to handle electrostatic interactions, with a short-range electrostatic interaction distance of 1.5 nm. Van der Waals interactions were calculated using a truncation method, with a cutoff value of 1.5 nm. The cold flow characteristic parameters of the natural ester insulating liquid were calculated using the equilibrium cross-flow autocorrelation function and the non-equilibrium periodic perturbation method. For example, the viscosity simulation results are shown in Table 3. By simulating the target properties of triglyceride molecules individually, it is convenient to construct the quantitative structure-property relationship of single molecules. This establishes a connection between the geometry of each molecule and its corresponding target properties, providing relevant simulation models and theoretical basis for the modification of natural ester molecules.

[0057] Table 3. Simulation results of kinematic viscosity of triglyceride molecules at different temperatures.

[0058]

[0059] 2) A quantitative fitting relationship between the cold flow characteristics of major molecules in natural ester insulating liquid and temperature was established based on the free volume theory, taking the kinematic viscosity characteristics of glycerol trioleate as an example. For simplification, the shear viscosity was approximated as the kinematic viscosity. According to the free volume theory, the relationship between kinematic viscosity, temperature, and free volume is shown in formula (1).

[0060]

[0061] In the formula, ν is the kinematic viscosity (cSt), T is the temperature (K), and V is the viscosity. f Free volume At the same temperature, A and B are constants. For ease of calculation, let B = 1. By substituting the molecular dynamics simulation value of the kinematic viscosity of trioleate, temperature, and free volume, the relationship between A and temperature can be calculated. Further, the relationship between free volume and temperature can be obtained. Substituting the results into formula (1), after simplification, the corresponding quantitative fitting relationship can be obtained:

[0062]

[0063] 3) Weak interactions between different natural ester molecules (hydrogen bonds, van der Waals interactions, electrostatic interactions, etc.) and within molecules (hydrogen bonds, dihydrogen bonds, van der Waals interactions, etc.) were investigated using IGM and RDG-based methods. Figure 4 The image shows the surface analysis diagram of the van der Waals potential of glycerol tristearate. The effects of altering the molecular structure characteristics of natural esters, such as characteristic functional groups, chain length, number of branches, branch length, and molecular saturation, on weak molecular interactions are analyzed. The results show that with increasing chain length, number of branches, and branch length, the weak interactions between triglyceride molecules are enhanced to some extent, thus affecting the cold flow properties of natural ester insulating liquids. This provides theoretical guidance for the modification of the molecular structure and performance improvement of natural esters.

[0064] 4) Based on the above-mentioned influencing factors, a molecular dynamics prediction model for the low-temperature kinematic viscosity of natural ester molecules with different proportions was constructed. The mixing ratios of four triglyceride molecules—glyceryl tristearate, glyceryl trioleate, glyceryl trilinoleate, and glyceryl trilinolenic acid—were varied and simulated. The results were compared with the cold flow characteristics of actual natural ester insulating liquids to obtain the molecular component mixing rules that can reduce the low-temperature kinematic viscosity and pour point of the natural ester insulating liquid system. The results show that compared with the common molecular ratio in natural ester insulating liquids (the ratio of glyceryl trilinolenic acid, glyceryl trioleate, glyceryl trioleate, and glyceryl tristearate is 10:50:24:16), reducing the proportions of glyceryl trioleate and glyceryl trilinolenic acid, and increasing the proportions of glyceryl tristearate and glyceryl trioleate, can improve the low-temperature kinematic viscosity and pour point characteristics of the natural ester insulating liquid to a certain extent. This indicates that the molecular geometry of glyceryl trioleate and glyceryl trilinolenic acid is closer to the ideal natural ester molecule modification results.

[0065] c) The study on the intrinsic molecular structure modification of natural ester insulating liquids based on quantitative structure-property relationships specifically includes the following steps:

[0066] 1) The integrated network-based molecular descriptor and fingerprinting platform, alvaDesc, was used to calculate molecular topological descriptors (based on molecular topology such as atomic distance, adjacency, and molecular shape, as well as molecular chemical information such as atomic charge and orbital information), quantum chemical property descriptors (based on density functional theory calculations), and physicochemical property descriptors (based on experimental data, literature, and data collected in previous work) for the quantitative structure-property relationship of natural ester molecules. Some descriptor calculation results are shown in Table 4. alvaDesc has advantages such as accurate and reliable results and a large number of computable molecular descriptors, capable of calculating over 5000 molecular descriptors across 33 categories. Furthermore, all molecular descriptor calculation algorithms provide different theoretical methods for calculating molecular descriptors of various triglyceride structures, offering strong interpretability.

[0067] Table 4. Calculation results of triglyceride molecular partial descriptors

[0068]

[0069]

[0070] 2) Construct a massive database of QSPR natural ester molecular descriptors based on SMILES molecular fingerprint characterization and precise 3D molecular structure characterization. The molecular descriptors are ordered by weight as follows: Figure 5 As shown. By Figure 5Taking the top three molecular descriptors by weight as examples, they are AMW, VE2sign_B(m), and rGes. AMW represents the average molecular weight of the reaction, belonging to the bulk index class of descriptors. VE2sign_B(m) represents the average coefficient of the last eigenvector of the mass-weighted burden matrix, belonging to the two-dimensional matrix-based descriptors. rGes represents the relative number of reaction symmetries (based on electrotopological states), belonging to the information index class of descriptors. This allows us to determine which molecular descriptors contribute significantly to the properties of natural ester molecules and guide molecular modification from the perspective of the corresponding descriptors.

[0071] 3) A high-throughput virtual screening study using computer-aided design was conducted to investigate the quantitative structure-property relationship between natural ester molecule descriptors and predicted properties. Principal component analysis and other feature selection methods, combined with neural network algorithms, were used to achieve preprocessing and efficient optimization of natural ester molecule descriptors. An example of the workflow is shown below. Figure 6 As shown, principal component analysis and recursive feature elimination were used to screen out a set of relevant molecular descriptors, and redundant descriptors were discarded. The optimized molecular descriptors were first passed through a Gaussian noise layer to achieve data augmentation. Then, they were input into a multi-layered neural network layer, and the output layer results were the properties of the target natural ester molecule. The rationality of the selected natural ester molecule descriptors was further verified by comparing the output results with the target properties. If the final error is within 10%, the constructed quantitative structure-property relationship can be considered reasonable, and the next step of molecular modification design can be carried out.

[0072] 4) Utilize tools such as molecular structure generators to add atoms or modify fragments to alter the initial conformation and descriptor probability distribution of natural ester molecules. Combined with molecular structural weak site analysis, quantify the correlation between molecular structural characteristic parameters and target properties. Use predictive models to redistribute and screen descriptors to achieve the desired targeted property molecular modification design and obtain the optimal modified natural ester molecule structure. Example structures are shown below. Figure 7 As shown.

[0073] The above embodiments are merely illustrative of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope thereof should be covered within the scope of the claims of the present invention.

Claims

1. A method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships, characterized in that, The natural ester molecules are triglyceride molecules, including triglyceride trilinoleate, triglyceride trilinoleate, triglyceride trioleate, triglyceride tristearate, and triglyceride tripalmitate; The process includes the following steps: a) establishing the correlation between the molecular structure and electrical properties of natural ester insulating liquids; b) establishing molecular dynamics predictions of the cold flow characteristics of natural esters and the influence of molecular configuration and components; c) performing intrinsic molecular structure modification of natural ester insulating liquids based on quantitative structure-property relationships. Establishing the correlation between the molecular structure and electrical properties of natural ester insulating liquids specifically includes the following steps: a.1) Optimize the molecular structure of natural ester insulating liquid molecules under the implicit solvent model at the computational level, and perform frequency vibration calculations to obtain a high-precision molecular structure without imaginary frequencies; a.2) Calculate the electronic structure energy / electronic excited state, and by analyzing the above calculation results, obtain the electronic structure property parameters and molecular structure geometric parameters that are strongly correlated with the molecular dielectric properties; a.3) Calculate the parameter changes after applying electric fields of different strengths, and analyze and compare the reasons and rules for the different dielectric properties of natural ester molecules due to different structural characteristics based on the parameter change law; a.4) Calculate the global index, real space function, and atomic index of the quantities reflecting the reactivity of chemical substances and the reaction sites in the concept density functional theory; The correlation between the molecular structure and electrical properties of natural ester insulating liquid was established.

2. The method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships according to claim 1, characterized in that, In step a.2), the electronic structure properties include ionization energy, electron affinity, electron density, dipole moment, average local ionization energy, polarizability, transition excitation energy, oscillator strength, band gap, nucleophilic and electrophilic hyperdelocalization electrical and photoelectric properties; the molecular structure geometric parameters include bond length, bond angle, dihedral angle, atomic type and number, and chemical bond type and number.

3. The method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships according to claim 1, characterized in that, Establishing molecular dynamics predictions and the influence of molecular configuration and components on the cold flow properties of natural esters specifically includes the following steps: b.1) Establish a molecular model of actual natural ester insulating liquid and study the simulation prediction of low-temperature kinematic viscosity and pour point of natural ester insulating liquid; b.2) Establish a quantitative fitting relationship between the cold flow characteristics of natural ester insulating liquid and temperature by combining the free volume theory; b.3) We will use IGM and RDG-based methods to investigate the weak interactions between and within different natural ester molecules and analyze the influence of changes in the molecular structure of natural esters on the weak interactions. b.4) Based on the above-mentioned influence laws, a molecular dynamics prediction model for the low-temperature kinematic viscosity of natural ester molecules with different proportions was constructed and compared with the cold flow characteristic parameters of actual natural ester insulating liquid. The study obtained the molecular component mixing rules that can reduce the low-temperature kinematic viscosity and pour point of the natural ester insulating liquid molecular system.

4. The method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships according to claim 3, characterized in that, In step b.1), the molecular dynamics calculation settings include using the OPLS-AA force field to describe the potential energy of the natural ester molecular system and the forces between each atom, using the Leap-frog method for dynamic simulation and setting periodic boundary conditions in all three directions of the system; using the Particle-Mesh Ewald method to handle electrostatic interactions, with a short-range electrostatic interaction distance of 1~2 nm; calculating van der Waals interactions using a truncation method, with a cutoff value of 1~2 nm; and using EMD and NEMD methods to calculate the cold flow characteristics parameters of the natural ester insulating liquid.

5. The method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships according to claim 3, characterized in that, In step b.3), the weak interactions between different natural ester molecules include hydrogen bonds, van der Waals interactions, and electrostatic interactions; the weak interactions within molecules include hydrogen bonds, dihydrogen bonds, and van der Waals interactions; and the molecular structural characteristics include characteristic functional groups, chain length, number of branches, branch length, and saturation.

6. The method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships according to claim 1, characterized in that, The intrinsic molecular structure modification of natural ester insulating liquid based on quantitative structure-property relationships specifically includes the following steps: c.1) Calculate the molecular topological descriptors, quantum chemical property descriptors, and physicochemical property descriptors in the quantitative structure-property relationship of natural ester molecules using an integrated network-based molecular descriptor and fingerprint computing platform; c.2) Construct a massive database of QSPR natural ester molecule descriptors based on SMILES molecular fingerprint characterization and precise 3D molecular structure characterization; c.3) The quantitative structure-property relationship between natural ester molecule descriptors and predicted properties was studied by computer-aided design of high-throughput virtual screening. The preprocessing and efficient optimization of natural ester molecule descriptors were achieved by combining statistical analysis and evolutionary algorithm methods. c.4) Use molecular structure generator tools to add atoms or modify fragments to change the initial conformation and descriptor probability distribution of natural ester molecules. Combine this with molecular structure weak site analysis to quantify the correlation between molecular structure feature parameters and target properties. Use a prediction model to redistribute and screen descriptors to achieve the desired targeted property molecular modification design and obtain the optimal modified natural ester molecule structure.

7. The method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships according to claim 6, characterized in that, In step c.1), the molecular topological descriptor includes atomic distances, adjacency relationships, molecular shape, and molecular chemical information, atomic electrical properties, and orbital information based on the molecular topological structure; the quantum chemical property descriptor is based on density functional theory calculations; and the physicochemical property descriptor is based on experimental data and collected from literature.

8. The method for modifying the molecular structure and improving the performance of natural esters based on quantitative structure-property relationships according to claim 6, characterized in that, In step c.3), the quantitative structure-property relationship construction method includes neural networks; the statistical analysis method includes principal component analysis and partial least squares; and the evolutionary algorithm includes genetic algorithm and differential evolution algorithm.

Images