A method for enhancing the nuclear magnetic resonance relaxation scale in molecular dynamics simulations

Abstract

The present invention discloses a method for enhancing the nuclear magnetic resonance relaxation scale in molecular dynamics simulations. The specific steps are as follows: S1: Use the molecular dynamics method to simulate the movement of fluid molecules at the solid-liquid boundary of porous media to obtain the longitudinal bulk relaxation T 1b and the transverse bulk relaxation T 2b ; S2: Obtain the longitudinal surface relaxation T 1S and the transverse surface relaxation T 2S ; S3: Calculate the equivalent longitudinal relaxation T 1eff and the equivalent transverse relaxation T 2eff at the solid-liquid boundary of porous media; S4: Use the two-layer lattice Boltzmann grid method to simulate the nuclear magnetic resonance relaxation characteristics, and transfer the obtained small-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary to a large-scale space, so as to obtain the large-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary. The present invention solves the scale matching problem in the application of the results of molecular dynamics simulations of nuclear magnetic resonance relaxation.

Description

Technical Field

[0001] The present invention belongs to the technical field of nuclear magnetic resonance relaxation in porous media, and particularly relates to a method for enhancing the nuclear magnetic resonance relaxation scale in molecular dynamics simulation. Background Art

[0002] Nuclear magnetic resonance relaxation technology (NMR) can obtain characteristic parameters such as porosity, permeability, and pore size distribution of porous media using relaxation time, and can also observe and analyze the dynamic behavior of fluid molecules. Therefore, people often use nuclear magnetic resonance relaxation simulation means to obtain the nuclear magnetic resonance response of porous media and extract pore information. The most commonly used is the Monte Carlo random walk method, which regards diffusing molecules as propagators and calculates the movement trajectories of the propagators. However, the Monte Carlo random walk method statistically calculates the average annihilation probability, which is a phenomenological theory. This method cannot handle the interactions between particles and is difficult to directly respond to surface adsorption and desorption in porous media. Therefore, it is necessary to upgrade NMR simulation from traditional Monte Carlo simulation to a method that considers molecular interactions and particle collisions.

[0003] With the enhancement of computer computing power and the improvement of force field models, molecular dynamics (MD) simulation has become a powerful tool for studying the structures and molecular dynamics of macromolecules such as proteins, polymers, and crude oil. MD simulation can distinguish the intramolecular hydrogen nuclear dipole-dipole interaction and the intermolecular dipole-dipole interaction involved in nuclear magnetic resonance relaxation, and explain the nuclear magnetic resonance relaxation mechanism of complex component fluids and porous media surfaces from the perspective of molecular interactions. This method has been successfully applied to the interpretation of the T1 and T2 relaxation mechanisms of crude oil and kerogen. However, the equivalent T1 and T2 obtained from MD simulation results can only represent the NMR characteristics within the boundary space at the nanometer (nm) scale, and the evolution time is at the picosecond (ps) level. Conventional NMR measurements cannot be simulated by such small-scale characteristics. Therefore, how to upgrade the MD simulation results to the NMR measurement scale so that the numerical simulation of NMR can simultaneously consider the changes in various parameters at the molecular level and pore level has become an important issue. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention discloses a method for enhancing the nuclear magnetic resonance relaxation scale in molecular dynamics simulation to solve the scale matching problem in the application of nuclear magnetic resonance relaxation results in molecular dynamics simulation.

[0005] The above technical object of the present invention is achieved by the following technical solutions:

[0006] A method for enhancing the nuclear magnetic resonance relaxation scale in molecular dynamics simulation, the specific steps are as follows:

[0007] S1: Use molecular dynamics method to simulate the movement of fluid molecules at the solid-liquid boundary of porous media, so as to obtain the translational diffusion characteristic time τ of fluid molecules t and rotational diffusion time τ r , calculate the longitudinal bulk relaxation T 1b and transverse bulk relaxation T 2b ;

[0008] S2: According to the adsorption and desorption processes of fluid molecules on the solid surface, count the surface diffusion time τ m and surface residence time τ s of fluid molecules on the solid surface, so as to calculate the longitudinal surface relaxation T 1S and transverse surface relaxation T 2S ;

[0009] S3: Calculate the equivalent longitudinal relaxation T 1eff and equivalent transverse relaxation T 2eff at the solid-liquid boundary of porous media respectively according to formulas (1) and (2), where formulas (1) and (2) are as follows:

[0010] T 1eff = 1 / (1 / T 1b +1 / T 1s ) (1);

[0011] T 2eff = 1 / (1 / T 2b +1 / T 2s ) (2);

[0012] S4: Use the equivalent longitudinal relaxation T 1eff and equivalent transverse relaxation T 2eff obtained by molecular dynamics simulation as the relaxation characteristics of the lattice at the small-scale boundary, use the two-layer lattice Boltzmann grid method to simulate the nuclear magnetic resonance relaxation characteristics, and transfer the obtained small-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary to the large-scale space, so as to obtain the large-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary.

[0013] Preferably, in S4, the specific steps of using the two-layer lattice Boltzmann grid method to simulate the nuclear magnetic resonance relaxation characteristics are as follows:

[0014] S4-1: Based on the equivalent longitudinal relaxation T 1eff and equivalent transverse relaxation T 2eff obtained by molecular dynamics simulation, use the Boltzmann grid method to simulate the nuclear magnetic resonance relaxation characteristics at a scale of 1000nm×1000nm×1000nm;

[0015] S4-2: Obtain the longitudinal relaxation distribution and transverse relaxation distribution at the nanoscale solid-liquid boundary through inversion calculation, and respectively take the maximum peak points of the longitudinal relaxation distribution and transverse relaxation as the equivalent longitudinal relaxation T1 and equivalent transverse relaxation T2 at the nanoscale solid-liquid boundary;

[0016] S4-3: Take 1 μm 3 as a single grid, take the equivalent longitudinal relaxation T1 and equivalent transverse relaxation T2 obtained in step S4-2 as the nuclear magnetic resonance relaxation characteristics of the grid at the solid-liquid boundary, and use the Boltzmann lattice method to simulate the nuclear magnetic resonance relaxation characteristics at the micron scale;

[0017] S4-4: Obtain the longitudinal relaxation distribution T1' and transverse relaxation distribution T2' at the micron scale through inversion calculation.

[0018] Preferably, the Boltzmann lattice method is the Shan-Chen LBM method, and the grid model used is the D3Q19 grid model.

[0019] Preferably, the specific method for simulating the nuclear magnetic resonance relaxation characteristics using the Boltzmann lattice method is as follows:

[0020] a Initialize the magnetization vector density of all grids;

[0021] b Calculate the equilibrium function according to the LBM collision and migration formula shown in formula (3), and the formula (3) is as follows:

[0022]

[0023] where ξ i is the discrete velocity, i represents the discrete velocity direction, x is the position coordinate of the function f, t is the current time step, f i (x,t) is the particle density distribution function of the fluid molecules along the velocity direction i at the spatial position x, δ t is the lattice time; τ velocity is the collision time, f i eq (x,t) is the equilibrium distribution function; among them, the discrete velocity ξ i follows the following formula:

[0024]

[0025] c Calculate the magnetization vector density of the collision and migration according to the equilibrium distribution function, multiply the magnetization vector density in each grid by the migration probability to perform the migration of the magnetization vector from this grid to the adjacent grid, and then sum it with the magnetization vector density of the fluid in the adjacent grid, and assign the sum result to the adjacent grid;

[0026] d Recalculate the sum of the magnetization vector densities of all cells within the current time step;

[0027] e Output the modulus of the sum of the magnetization vector densities of all cells as the echo signal according to the CPMG pulse sequence;

[0028] f If the echo signal decays to 0.001 times the maximum value, stop the loop;

[0029] g Invert the echo signal to obtain the overall equivalent transverse relaxation distribution and equivalent longitudinal relaxation distribution.

[0030] Preferably, the two - layer Boltzmann lattice method simulation can evolve independently.

[0031] Beneficial effects: The present invention discloses a method for enhancing the nuclear magnetic resonance relaxation scale of molecular dynamics simulation, having the following advantages:

[0032] (1) The present invention adopts the double - layer Boltzmann (LBM) lattice method, which can simply and effectively upscale the molecular dynamics (MD) simulation results to the scale verifiable by nuclear magnetic resonance physical experiments, providing a more accurate and convenient numerical experimental method for the nuclear magnetic resonance relaxation simulation of porous media.

[0033] (2) By the independent operation of cells at different scales, the present invention reduces a large amount of data exchange, which is beneficial to improving the efficiency of computer parallel computing.

[0034] (3) The present invention combines molecular dynamics (MD) simulation and nuclear magnetic resonance measurement scale. Compared with traditional nuclear magnetic resonance simulations, it considers molecular interactions and particle collisions, can directly respond to surface adsorption and desorption in porous media, and enables macroscopic nuclear magnetic resonance phenomena to be explained by microscopic simulations. Description of the Drawings

[0035] Figure 1 is the implementation flowchart of the present invention;

[0036] Figure 2 is the comparison diagram of the micron - scale transverse relaxation distribution T2′ obtained by three simulation methods in Example 1 (200 - nm - diameter small balls).

[0037] Figure 3 is the comparison diagram of the micron - scale transverse relaxation distribution T2′ obtained by three simulation methods in Example 2 (1 - μm - diameter small balls). Detailed Embodiments

[0038] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0039] Example 1

[0040] Taking silica as the porous medium and water as the fluid as the simulation object, this example provides a method for enhancing the nuclear magnetic resonance relaxation scale of molecular dynamics simulation. The specific steps are as follows:

[0041] S1: Using a 200-nm diameter sphere as a model, the movement of water molecules at the solid-liquid boundary of the porous medium is simulated by molecular dynamics to obtain the translational diffusion characteristic time τ of water molecules t and the rotational diffusion time τ r , and calculate the bulk relaxation T 1b and T 2b ;

[0042] S2: During the adsorption and desorption processes of fluid molecules on the solid surface, the surface distance between the fluid molecules and the solid surface is selected according to the range involved in the atomic force field of the solid surface at the boundary in molecular dynamics simulation. In this example, for water molecules and the silica surface, when the surface distance between the two is ≥ 1 nm, it is considered that the water molecules exchange with the bulk water molecules and leave the surface. Therefore, the surface diffusion time τ m and the surface residence time τ s of water molecules are respectively statistically analyzed, so as to calculate the longitudinal surface relaxation T 1S and the transverse surface relaxation T 2S ;

[0043] S3: According to formulas (1) and (2), the equivalent longitudinal relaxation T 1eff and the equivalent transverse relaxation T 2eff at the solid-liquid boundary of the porous medium are respectively calculated, where formulas (1) and (2) are as follows:

[0044] T 1eff = 1 / (1 / T 1b +1 / T 1s ) (1);

[0045] T 2eff = 1 / (1 / T 2b +1 / T 2s ) (2);

[0046] S4: Taking the equivalent longitudinal relaxation T 1eff and the equivalent transverse relaxation T 2eff obtained by molecular dynamics simulation as the relaxation characteristics of the lattice at the small-scale boundary, the two-layer lattice Boltzmann grid method is used to simulate the nuclear magnetic resonance relaxation characteristics, and the small-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary are transferred to the large-scale space, so as to obtain the large-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary. The specific method is as follows:

[0047] S4-1: Based on the equivalent longitudinal relaxation T 1eff and the equivalent transverse relaxation T 2eff , the nuclear magnetic resonance relaxation characteristics at the scale of 1000nm×1000nm×1000nm are simulated using the Boltzmann lattice Boltzmann method;

[0048] S4-2: The longitudinal relaxation distribution and the transverse relaxation distribution at the solid-liquid boundary at the nanoscale are obtained through inversion calculation, and the maximum peak points of the longitudinal relaxation distribution and the transverse relaxation are respectively taken as the equivalent longitudinal relaxation T1 and the equivalent transverse relaxation T2 at the solid-liquid boundary at the nanoscale;

[0049] S4-3: Taking 1μm 3 as a single lattice, and taking the equivalent longitudinal relaxation T1 and the equivalent transverse relaxation T2 obtained in step S4-2 as the nuclear magnetic resonance relaxation characteristics of the lattice at the solid-liquid boundary, the nuclear magnetic resonance relaxation characteristics at the micron scale are simulated using the Boltzmann lattice Boltzmann method;

[0050] S4-4: The longitudinal relaxation distribution T1' and the transverse relaxation distribution T2' at the micron scale are obtained through inversion calculation.

[0051] In this embodiment, the Boltzmann lattice Boltzmann method used in S4-1 and S4-2 is the Shan-Chen LBM method, and the grid model used is the D3Q19 grid model.

[0052] In this embodiment, the specific method for simulating the nuclear magnetic resonance relaxation characteristics by the LBM grid method is as follows:

[0053] a Initialize the magnetization vector density of all lattices;

[0054] b Calculate the equilibrium function according to the LBM collision and migration formula shown in formula (3), and the formula (3) is as follows:

[0055]

[0056] where ξ i is the discrete velocity, i represents the discrete velocity direction, x is the position coordinate of the function f, t is the current time step, f i (x,t) is the particle density distribution function of the fluid molecules along the velocity direction i at the spatial position x, δ t is the lattice time; τ velocity is the collision time, f i eq (x,t) is the equilibrium distribution function; among them, the discrete velocity ξ i follows the following formula:

[0057]

[0058] c Calculate the magnetization vector density of collision migration according to the equilibrium distribution function, multiply the magnetization vector density in each grid by the migration probability to migrate the magnetization vector of this grid to the adjacent grid, then sum it with the magnetization vector density of the fluid in the adjacent grid, and assign the sum result to the adjacent grid;

[0059] d Recalculate the sum of the magnetization vector densities of all grids within this time step;

[0060] e Output the modulus of the sum of the magnetization vector densities of all grids as the echo signal according to the CPMG pulse sequence;

[0061] f If the echo signal decays to 0.001 times the maximum value, stop the loop;

[0062] g Invert the echo signal to obtain the overall equivalent transverse relaxation distribution and equivalent longitudinal relaxation distribution.

[0063] In this embodiment, the two - layer Boltzmann lattice method simulation can evolve independently.

[0064] For the simulation object of this Embodiment 1 (using a 200 - nm - diameter sphere as the model), the Monte Carlo random walk method and the LBM direct simulation method are respectively used for nuclear magnetic resonance relaxation simulation to obtain the relaxation distribution at the micron scale of the corresponding porous medium (silica). As Figure 2 Shown is the transverse relaxation distribution T′2 at the micron scale of the porous medium finally obtained by the three methods. It can be seen from the figure that the nuclear magnetic resonance relaxation simulation results based on the LBM upscaling in this Embodiment 1 basically coincide with the simulation results of the existing Monte Carlo random walk method and the LBM direct simulation method, indicating that the simulation method in this Embodiment 1 is feasible and the obtained simulation results are reliable. Compared with the existing Monte Carlo random walk method and the LBM direct simulation method that need to obtain the simulation boundary conditions through assumptions, this Embodiment 1 can generate the boundary conditions required for nuclear magnetic resonance simulation through molecular dynamics simulation: the equivalent relaxation time on the solid - liquid boundary surface, making the simulation calculation have a more theoretical basis at the molecular scale. In addition, the obtained relaxation time can be transferred at different scales, which greatly facilitates the nuclear magnetic resonance simulation work of porous media.

[0065] Embodiment 2

[0066] Taking silica as the porous medium and water as the fluid as the simulation object, and using a 1 - μm - diameter sphere as the model, nuclear magnetic resonance relaxation simulation is carried out based on molecular dynamics simulation and the two - layer Boltzmann lattice method. The specific method is the same as that in Embodiment 1.

[0067] At the same time, the Monte Carlo random walk method and the LBM direct simulation method are used for nuclear magnetic resonance relaxation simulation to carry out nuclear magnetic resonance relaxation simulation on the above - mentioned simulation object.

[0068] As Figure 3 shown are the transverse relaxation distributions T2' at the microscale of the porous medium finally obtained by three methods. It can be seen from the figure that the simulation results of nuclear magnetic resonance relaxation based on the LBM upscaling in this Embodiment 2 basically coincide with the simulation results of the existing Monte Carlo random walk method and the LBM direct simulation method, indicating that the simulation method in this Embodiment 2 is feasible and the obtained simulation results are reliable. This specific embodiment is only an explanation of the present invention and is not a limitation thereof. Those skilled in the art can make modifications without creative contributions to this embodiment according to needs after reading this specification, but as long as it is within the scope of the claims of the present invention, it is protected by the patent law.

Claims

1. A method for enhancing the nuclear magnetic resonance relaxation scale of molecular dynamics simulations, characterized in that, The specific steps are as follows: S1: Simulate the motion of fluid molecules at the solid-liquid boundary of porous media using molecular dynamics methods to obtain the translational diffusion characteristic time τ of fluid molecules t and the rotational diffusion time τ r , calculate the longitudinal bulk relaxation T 1b and the transverse bulk relaxation T 2b ; S2: According to the adsorption and desorption processes of fluid molecules on the solid surface, statistically calculate the surface diffusion time τ of fluid molecules on the solid surface m and the surface residence time τ s , thereby calculating the longitudinal surface relaxation T 1S and the transverse surface relaxation T 2S ; S3: Calculate the equivalent longitudinal relaxation T at the solid-liquid boundary of the porous medium according to formulas (1) and (2) respectively 1eff and the equivalent transverse relaxation T 2eff , where formulas (1) and (2) are as follows: T 1eff = 1 / (1 / T 1b + 1 / T 1s )(1); T 2eff = 1 / (1 / T 2b + 1 / T 2s )(2); S4: Use the equivalent longitudinal relaxation T obtained from molecular dynamics simulation 1eff and the equivalent transverse relaxation T 2eff as the relaxation characteristics of the lattice at the small-scale boundary. Use the two-layer lattice Boltzmann grid method to simulate the nuclear magnetic resonance relaxation characteristics, and transfer the obtained small-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary to the large-scale space, so as to obtain the large-scale nuclear magnetic resonance relaxation characteristics at the solid-liquid boundary; In S4, the specific steps of using the two - layer lattice Boltzmann grid method to simulate the nuclear magnetic resonance relaxation characteristics are as follows: S4-1: Equivalent longitudinal relaxation T obtained based on molecular dynamics simulation 1eff and equivalent transverse relaxation T 2eff , and the nuclear magnetic resonance relaxation characteristics at a scale of 1000 nm × 1000 nm × 1000 nm are simulated using the Boltzmann grid method; S4 - 2: Obtain the longitudinal relaxation distribution and transverse relaxation distribution at the nano - scale solid - liquid boundary through inversion calculation, and take the maximum peak points of the longitudinal relaxation distribution and transverse relaxation respectively as the equivalent longitudinal relaxation T1 and equivalent transverse relaxation T2 at the nano - scale solid - liquid boundary; S4-3: Take 1 μm 3 As a single grid, use the equivalent longitudinal relaxation T1 and equivalent transverse relaxation T2 obtained in step S4-2 as the nuclear magnetic resonance relaxation characteristics of the grid at the solid-liquid boundary, and simulate the nuclear magnetic resonance relaxation characteristics at the micron scale using the Boltzmann grid method; S4 - 4: Obtain the longitudinal relaxation distribution T1' and transverse relaxation distribution T2' at the micro - scale through inversion calculation.

2. The method for enhancing the nuclear magnetic resonance relaxation scale of molecular dynamics simulation according to claim 1, wherein The Boltzmann grid method is the Shan - Chen LBM method, and the grid model used is the D3Q19 grid model.

3. The method for enhancing the nuclear magnetic resonance relaxation scale of molecular dynamics simulation according to claim 1, wherein The specific method of using the Boltzmann grid method to simulate the nuclear magnetic resonance relaxation characteristics is as follows: a Initialize the magnetization vector density of all grids; b Calculate the equilibrium function according to the LBM collision and migration formula shown in formula (3), and formula (3) is as follows: Among them, ξ i is the discrete velocity, i represents the discrete velocity direction, x is the position coordinate of the function f, t is the current time step, and f i (x, t) is the particle density distribution function of fluid molecules at the spatial position x along the velocity direction i, and δ t is the lattice time; τ velocity is the collision time, and f i eq (x, t) is the equilibrium distribution function; among them, the discrete velocity ξ i follows the following formula: c Calculate the magnetization vector density of collision and migration according to the equilibrium distribution function, multiply the magnetization vector density in each grid by the migration probability to migrate the magnetization vector of this grid to the adjacent grid, then sum it with the magnetization vector density of the fluid in the adjacent grid, and assign the sum result to the adjacent grid; d Recalculate the sum of the magnetization vector densities of all grids within the current time step; e Output the modulus of the sum of the magnetization vector densities of all grids as the echo signal according to the CPMG pulse sequence; f If the echo signal decays to 0.001 times the maximum value, stop the loop; g Invert the echo signal to obtain the overall equivalent transverse relaxation distribution and equivalent longitudinal relaxation distribution.

4. The method for enhancing the nuclear magnetic resonance relaxation scale of molecular dynamics simulation according to claim 1, wherein, The two - layer Boltzmann grid method simulation can evolve independently respectively.

Images