A method for predicting the separation performance of a selectively permeable membrane

Abstract

The application discloses a kind of selective permeable membrane separation performance prediction method, comprising the following steps: constructing membrane separation model, including separation process module, channel model module and mass transfer module;Obtain the data parameters and model architecture of the membrane separation experiment needed to simulate, and construct the model architecture of the currently simulated membrane separation experiment in the membrane separation model;Obtain the model equation under the model architecture, introduce trapezoidal rule, correlate the relationship between the flow change of the permeation side and the flow change of the retentate side, based on the component permeation flux, finally can form the industrial scale level membrane area, separation coefficient, terminal permeation flow, terminal permeation pressure, terminal retentate flow, terminal retentate pressure, concentration distribution, velocity distribution and pressure distribution etc. All-around simulation result display, all aspects that cannot be simultaneously paid attention to in experiment are presented, compared with other models, model result is more comprehensive and has reference.

Description

Technical Field

[0001] This application relates to the field of membrane separation technology, and in particular to a method for predicting the performance of selective permeation membrane separation. Background Technology

[0002] Membrane separation is the process of separating gas or liquid mixtures using membrane materials. The separation process generally involves three streams: a feed stream, a permeate stream, and a tail stream that is in equilibrium with the feed and permeate streams. The main component of membrane separation is a tubular membrane filament formed from selective membrane materials using specific preparation and processing methods. The structure and size of the membrane filament affect the membrane separation process. In actual separation processes, to achieve a certain separation effect, one or more membrane filaments can be arranged in parallel within the membrane device. Membrane separation is widely used in environmental remediation and industrial energy conservation; however, a computational model that can quickly and accurately describe the membrane separation process is currently lacking.

[0003] Current models suffer from shortcomings in functionality, performance efficiency, ease of use, reliability, and portability, failing to meet the requirements of industrial applications. Functionally, existing models exhibit limited applicability, often only suitable for calculations of single systems or processes, and cannot meet accuracy requirements for multi-component systems or complex processes. In terms of performance efficiency, they consume significant amounts of computational memory, have long computation times, and slow response speeds. User-friendly operation prompts are difficult for users to understand. Reliability is unstable, prone to issues such as prolonged inaccessibility and system crashes. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method for predicting the performance of selective permeation membrane separation. This method can accurately predict the separation performance of a selective permeation membrane separation device and obtain the apparent permeability and selectivity of the membrane material, thereby enabling the verification of the membrane separation process and meeting the needs of membrane separation prediction in industrial applications.

[0005] The selective permeation membrane separation performance prediction method provided in this embodiment includes the following steps:

[0006] Step S1: Construct a membrane separation model based on the dynamic relationship between feed flow and permeate flow in membrane separation;

[0007] Step S2: Obtain the data parameters and model architecture of the membrane separation to be simulated, and construct the model architecture of the currently simulated membrane separation in the membrane separation model; wherein, the input variables of the data parameters of the membrane separation to be simulated include at least the feed flow rate, the concentration of each component, the permeation rate of each component, the structural parameters of the membrane module, the operating pressure and temperature;

[0008] Step S3: Based on the constructed model architecture, generate simulation data, obtain the model equations under the model architecture, and correlate the flow rate changes on the permeate side and the flow rate changes on the residual side during the solution process of the model equations to quickly obtain the changes in permeate flow rate, residual flow rate, tail gas, concentration of each component, and pressure of each component.

[0009] Furthermore, in step S3), the trapezoidal rule is introduced into the solution process of the model equation to optimize the solution steps:

[0010] Based on the trapezoidal rule, the integral interval in the model equation is divided into several trapezoidal infinitesimal elements to obtain the area under the curve. By constructing an operator in each infinitesimal element in which the decrease in permeable flow is equal to the increase in residual flow, and introducing the operator into the model equation under the model framework, the model equation is transformed. Under the constraint of boundary conditions, the changes in permeable flow, residual flow, concentration of each component, and pressure of each component are quickly obtained.

[0011] Furthermore, the step of dividing the integration interval in the model equation into several trapezoidal infinitesimal elements based on the trapezoidal rule to obtain the area under the curve specifically includes:

[0012] Based on the trapezoidal rule, the model equation f(x) is transformed over the integration interval [a,b]:

[0013]

[0014] Transformation of model equations based on n equidistant trapezoidal infinitesimal elements:

[0015] Where x0 = a,x n =b.

[0016] Furthermore, taking the pressure change along the length of the membrane tube as an example, the equation is:

[0017]

[0018] This invention discretizes it into n equidistant subintervals on [0,L]: z0=0, z1=Δz, z2=2Δz,…,z n =L, where Δz = L / n, and transform the integral equation into a trapezoidal algebraic equation over n equidistant subintervals, thus enabling the equation to be solved:

[0019]

[0020] Further, in step S1):

[0021] The membrane separation model includes a separation process module, a channel model module, and a mass transfer module.

[0022] The separation process module is constructed based on the flow directions of the feed stream and the permeate stream;

[0023] The channel model module is constructed by selecting the corresponding flow channel based on the direction of the raw material flow;

[0024] The mass transfer module is constructed based on the transfer relationship of substances across the membrane, the variation relationship along the length of the membrane module, and whether concentration polarization occurs in the membrane separation model.

[0025] Furthermore, the separation process module includes: a counter-current unit, a parallel-current unit, and a cross-current unit;

[0026] In the countercurrent unit, the flow directions of the feed stream and the permeate stream on both sides of the membrane are opposite;

[0027] In the parallel flow unit, the feed stream and the permeate stream on both sides of the membrane flow in the same direction;

[0028] In the cross-flow unit, the flow directions of the feed stream and the permeate stream on both sides of the membrane are intersected. In the approximation, it is assumed that the two flow directions are the same and that there is a difference in concentration boundaries.

[0029] Furthermore, the channel model module includes: tube side units and shell side units;

[0030] In the tube-side unit, easily permeable substances in the feed stream permeate from inside the membrane fibers to the shell side of the membrane separation device;

[0031] In the shell-side unit, easily permeable substances in the feed stream permeate from the outside of the membrane fibers to the inside of the membrane fibers;

[0032] In the channel model module, a total mass conservation model, a material conservation model, and a pressure model are constructed.

[0033] Furthermore, the mass transfer module includes: an interface transfer module and a convection diffusion module;

[0034] The interface transfer module is based on a model of mass transfer across the membrane, and constructs a permeation rate model for each component based on the relationship between feed flow rate, permeate flow rate, and residual flow rate.

[0035] When the difference in permeation rates of each component does not exceed a preset value, no concentration polarization phenomenon occurs in the membrane separation model. Then, the convection diffusion module constructs the permeation model and boundary conditions for each component based on the relationship between the concentration gradient, pressure gradient and velocity gradient along the length of the membrane module.

[0036] When the difference in permeation rates among the components exceeds a preset value, concentration polarization occurs in the membrane separation model. Then, the convection diffusion module introduces a concentration polarization model based on the concentration gradient, pressure gradient, velocity gradient, and concentration polarization phenomenon along the length of the membrane module to construct the permeation model of each component, as well as the boundary conditions, mass transfer driving force of the concentration polarization layer, and molar gas concentration of the components within the concentration polarization layer.

[0037] Furthermore, when concentration polarization occurs in the membrane separation model, the corresponding interface transfer equation is: d(ux)=k(x-xs)+xdu, where,

[0038] Where u represents the fluid flow rate on the permeate side of the membrane module, v represents the fluid flow rate on the permeate side of the membrane module, x represents the molar concentration on the permeate side of the membrane module, y represents the molar concentration on the permeate side of the membrane module, and p i J1 represents the pressure inside the fiber, J2 is the permeability coefficient, and k is the mass transfer coefficient.

[0039] The interface transfer equations are optimized and solved using the trapezoidal rule, including:

[0040] Step 1: Discretize the integration interval [0, L] into n subintervals: z0 = 0, z1 = Δz, z2 = 2Δz, ..., z n =L, where Δz = L / n.

[0041] Step 2: Establish an alternative equation for the integral equation based on the trapezoidal rule:

[0042]

[0043] Step 3: Apply the substitution equation to solve the interface transfer equation.

[0044]

[0045] definition:

[0046] Establish an alternative equation for solving the interface transfer equation based on the trapezoidal rule:

[0047]

[0048] Step 4: Calculate f(z) i ): Calculate each x(z) i ) and x s (z i ), and thus obtain f(z) i ).

[0049] Step 5: Iterative solution. Based on each f(z) iThe value is iterated and the distance z is updated. i Permeation flux at point ux(z) i ).

[0050] Furthermore, assuming the calculation interval of the membrane tube with length [0, L] is divided into 4 sub-intervals, the interface transfer equation is solved.

[0051] Step 1: z0 = 0, z1 = L / 4, z2 = L / 2, z3 = 3L / 4, z4 = L.

[0052] Step 2: Calculate x(z0) and x s (z0) value.

[0053] Step 3: According to Calculate the value of f(z0) and update the value of ux(z1).

[0054] Step 4: Based on the updated value of ux(z1), iteratively calculate f(z1). Repeat this process to obtain f(z2), f(z3), and f(z4) respectively.

[0055] Step 5: The final interface throughput is:

[0056] Actual interface saturation concentration x s The expression is very complex:

[0057] The integral form of the interface transfer equation is:

[0058]

[0059] Using the solution method in steps 1 to 5 can improve the solution speed of the equation.

[0060] Furthermore, the selective permeation membrane separation performance prediction method also includes:

[0061] Step S4: Obtain the prediction results and corresponding real data from each prediction, perform polynomial fitting on the error between the two to obtain the error correction value, and then... Train the output neural network model, where All are matrices, and the polynomial fitting is adjusted to obtain the final prediction correction model.

[0062] Furthermore, the variation of the membrane separation model in the length direction is calculated using the following equation:

[0063] The equation for calculating the residual flow rate u is: u(i+1)=u(i)-du, where u(i) is the flow rate at the i-th node on the residual side;

[0064] The equation for calculating the infiltration flow rate v is: v(i+1)=v(i)+dv, where v(i) is the flow rate at the i-th node on the infiltration side;

[0065] The equation for calculating the osmotic concentration x is: x(i+1)=x(i)-dx, where x(i) is the concentration of the component at the i-th node on the osmotic side;

[0066] The equation for calculating the osmotic concentration y is: y(i+1)=y(i)+dy, where y(i) is the component concentration at node i on the osmotic side.

[0067] Furthermore, the membrane separation model has at least two components;

[0068] The variation of any component j is calculated using the following model:

[0069] Permeate flow rate u j Calculation equation: u j (i+1)=u j (i)-du;

[0070] Infiltration flow v j Calculation equation: v j (i+1)=v j (i)+dv;

[0071] residual concentration x j Calculate the equation: x j (i+1)=x j (i)-dx;

[0072] Osmotic concentration y j Calculate the equation: y j (i+1)=y j (i)+dy;

[0073] The total pressure is the sum of the pressures of each component: m represents the number of components in the system;

[0074] The total residual flow rate u is the sum of the residual flow rates of each component:

[0075] The total residual flow rate v is the sum of the residual flow rates of each component:

[0076] The sum of the residual concentrations calculated as percentages is 1:

[0077] The sum of the osmotic concentrations calculated as percentages is 1:

[0078] The beneficial effects of this invention are:

[0079] 1) The membrane separation model constructed in this application simulates membrane separation processes for industrialization (including pilot and industrial scale). Other existing models are mainly for chemical processes. The two have great differences in interfacial permeation performance. The prediction scale of other models is usually at the centimeter, millimeter and below, while the prediction scale of this model is at the meter and above.

[0080] 2) The model in this application is based on the component permeation flux, and can ultimately generate a comprehensive simulation result display of membrane area, separation coefficient, terminal permeation flow rate, terminal permeation pressure, terminal residual flow rate, terminal residual pressure, concentration distribution, velocity distribution and pressure distribution at the industrial scale. It presents all aspects that cannot be considered simultaneously in the experiment. Compared with other models, the model results are more comprehensive and have reference value. At the same time, for membrane separation processes under various application scenarios, the relative deviation of the data obtained by this solution can be less than 20%, and the data reliability is high.

[0081] 3) The simulation time for each step of the membrane separation process can be kept at around 20 milliseconds, achieving efficient and fast calculation. Attached Figure Description

[0082] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application, and other drawings obtained on this basis should also be attributed to this embodiment.

[0083] Figure 1 This is a flowchart of the selective permeation membrane separation performance prediction method of this application;

[0084] Figure 2 This is a framework diagram of the membrane separation model in this application;

[0085] Figure 3 This is a flowchart of the reverse flow unit in this application;

[0086] Figure 4 This is a flowchart illustrating the parallel flow unit in this application;

[0087] Figure 5 This is a flowchart of the cross-flow unit in this application. Detailed Implementation

[0088] To make the purpose, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0089] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0090] In the description of this application, it should be understood that the terms "upper", "lower", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0091] The present application will be described below with reference to specific embodiments:

[0092] The model of this invention is designed for industrial processes, namely membrane separation processes for industrialization (including pilot and industrial scale). Under this scheme, the scale of membrane separation is usually at the meter level and above.

[0093] To address the multifaceted calculations required for experimental data in this model environment, this application proposes a method for predicting the separation performance of large-scale membrane separation devices, comprising the following steps:

[0094] Step S1: Construct a membrane separation model based on the dynamic relationship between feed flow and permeate flow in membrane separation. This model is designed for industrial processes. Figure 2 As shown, this model is divided into three levels, including the separation process module, the channel model module, and the mass transfer module.

[0095] The separation process module involves the selection of membrane separation processes. Membrane separation processes generally have three modes: countercurrent, cocurrent, and crosscurrent. All three modes will appear in engineering practice. When building the model, countercurrent, cocurrent, and crosscurrent process modes are provided for selection.

[0096] like Figure 2 As shown in the figure, as a specific embodiment, the separation process module includes: a countercurrent unit, a cocurrent unit, and a cross-flow unit. In the countercurrent unit, the flow directions of the feed stream and the permeate stream on both sides of the membrane are opposite. In the cocurrent unit, the flow directions of the feed stream and the permeate stream on both sides of the membrane are the same. In the cross-flow unit, the flow directions of the feed stream and the permeate stream on both sides of the membrane are interleaved. In approximation, it is assumed that the two flow directions are the same and that there is a difference in concentration boundaries.

[0097] As a specific embodiment, a hollow fiber membrane module is used as a membrane separation device in industrial applications for illustration:

[0098] like Figure 3 The diagram shows a flow chart of the countercurrent unit, in which the feed flow and permeate flow on both sides of the hollow fiber membrane flow in opposite directions.

[0099] like Figure 4 The schematic diagram of the co-current unit shown shows that the feed flow and permeate flow on both sides of the hollow fiber membrane have the same flow direction.

[0100] like Figure 5 The schematic diagram of the cross-flow unit shown in the figure indicates that the flow directions of the feed flow and the permeate flow on both sides of the hollow fiber membrane are considered to be the same, and the concentration boundaries are different.

[0101] In the above diagram, u represents the fluid flow rate on the permeate side of the membrane module, v represents the fluid flow rate on the permeate side of the membrane module, x represents the molar concentration on the permeate side of the membrane module, and y represents the molar concentration on the permeate side of the membrane module.

[0102] The channel model module constructs a path for selecting the material flow channel, offering two options: tube-side or shell-side. Tube-side flow means that easily permeable substances in the material flow permeate from the inside of the membrane fibers to the shell side of the membrane separation unit. Shell-side flow means the permeation direction is reversed, with easily permeable substances permeating from the outside of the membrane fibers to the inside. From a material transport perspective, the two flow methods are essentially the same, differing only in the pressure drop.

[0103] like Figure 2 As shown in the figure, as a specific embodiment, the channel model module includes: tube side unit and shell side unit.

[0104] In the tube-side unit, easily permeable substances in the feed stream permeate from the inside of the membrane fibers into the shell side of the membrane separation unit.

[0105] In the shell-side unit, easily permeable substances in the feed stream permeate from the outside of the membrane fibers to the inside of the membrane fibers.

[0106] In the channel model module, tube-side or shell-side units are selected according to the actual membrane separation situation to construct the total mass conservation model, material conservation model and pressure model.

[0107] As a specific example, taking the raw material flow as a two-component model, the conservation model equations constructed in the tube unit of the channel model module are as follows:

[0108] Total mass conservation equation: du = dv;

[0109] Material conservation equation for component 1: udx + xdu = vdy + ydv;

[0110] Material conservation equation for component 2: -udx + (1-x)dv = -vdy + (1-y)dv;

[0111] Construct the pressure equation along the radial direction z:

[0112] Where μ represents the viscosity of the mixture, p i d represents the internal pressure of the wire. i The inner diameter of the hollow fiber membrane is represented by R, the gas constant is T, the temperature is N, and the number of membrane fibers in the membrane separation device is N.

[0113] As a specific example, taking the feedstock flow as a two-component model, the conservation model equations constructed in the shell-side element of the channel model module are as follows:

[0114] Construct the conservation model equations:

[0115] Total mass conservation equation: du = dv;

[0116] Material conservation equation for component 1: udx + xdu = vdy + ydv;

[0117] Material conservation equation for component 2: -udx + (1-x)dv = -vdy + (1-y)dv;

[0118] Construct the pressure equation along the radial direction z:

[0119] Where S represents the average fluid velocity on the shell side, p is the pressure, D represents the equivalent diameter of the membrane filament, D0 is the outer diameter of the membrane filament, and N is the number of membrane filaments in the membrane separation device.

[0120] The mass transfer module is constructed based on the transfer relationship of substances across the membrane, the variation relationship along the length of the membrane module, and whether concentration polarization occurs in the membrane separation model.

[0121] Specifically, the mass transfer module includes an interface transfer module and a convection diffusion module. The interface transfer module is a model for mass transfer across the membrane, generally (the difference in component permeation rates can be limited to a preset value, which is selected based on empirical values). The convection diffusion model includes models of concentration gradients, pressure gradients, and velocity gradients along the length of the membrane module. When the difference in component permeation rates exceeds the preset value, the influence of concentration polarization on the interface concentration distribution needs to be considered in the convection diffusion module. Here, concentration polarization is a special transfer phenomenon caused by the surface properties of the membrane material. That is, when the concentration difference between easily permeable and difficult-to-permeate components on the membrane surface is extremely large (<10%, which is taken as the aforementioned empirical value), the difficult-to-permeate component reacts to the easily permeable component, resulting in a decrease in the mass transfer capacity of the easily permeable component.

[0122] As a specific implementation, the state equation is constructed in the interface delivery module:

[0123] Among them, u in Indicates the feed flow rate, v out Indicates the permeation flow, u out x represents the residual flow rate. out Indicates the residual composition, y out Indicates the osmotic composition, J A J represents the permeability coefficient of component 1. B Q represents the permeability coefficient of component 2. A Q is the permeation rate of component 1. B x represents the permeation rate of component 2. in p1 represents the feed concentration, p2 represents the permeate side pressure, and p2 represents the permeate side pressure.

[0124] As a specific embodiment, when the difference in permeation rates of each component does not exceed a preset value, no concentration polarization phenomenon occurs in the membrane separation model. Under this condition, the convection diffusion module constructs the permeation model and boundary conditions for each component based on the relationship between the concentration gradient, pressure gradient, and velocity gradient along the length of the membrane module.

[0125] As a specific example, the state equations are constructed in the convection-diffusion model:

[0126] Component 1 Permeability Equation: J A (p i xp o y)dz / L=-udx-xdu;

[0127] Component 2 Permeability Equation: J A [p i (1-x)-p o (1-y)]dz / (Lα)=udx-(1-x)du;

[0128] Among them, the permeability coefficient J varies with temperature and follows the Arrhenius relation;

[0129] Permeability coefficient:

[0130] Separation coefficient α = J1 / J2;

[0131] Mixture viscosity:

[0132] Boundary conditions:

[0133] At z = 0, x = x0, u = u0, p i =p i,0 ,

[0134] At z = L, v = 0, p o =p o ,

[0135] p o p represents the external pressure on the wire. i The formula represents the internal pressure of the filament, L represents the filament length, E1 represents the activation energy, C1 represents a constant, J2 is the permeability coefficient of the difficult-to-permeate component, x0 is the feed concentration, u0 is the feed flow rate, and p i,0 For the feed pressure, x L Let y be the effluent concentration on the permeate side at a location with a membrane tube length of L. L The concentration of the permeate fluid at a position with a membrane tube length of L;

[0136] When the difference in permeation rates among the components exceeds a preset value, concentration polarization occurs in the membrane separation model. Under this condition, the convection diffusion module introduces concentration polarization to construct a model based on the concentration gradient, pressure gradient, velocity gradient, and concentration polarization phenomenon along the length of the membrane module. This model constructs the permeation model for each component, as well as the boundary conditions, mass transfer driving force of the concentration polarization layer, and molar gas concentration of the components within the concentration polarization layer.

[0137] As a specific implementation, when the difference in component permeation rates exceeds a preset value, a concentration polarization model is introduced into the convection-diffusion model to construct the equation of state:

[0138] Component 1 Permeability Equation: J A (p i x s -p o y)dz / L=-udx-xdu;

[0139] Component 2 Permeability Equation: J A [p i (1-x s )-p o(1-y)]dz / (Lα)=udx-(1-x)du;

[0140] Concentration polarization mass transfer driving force: d(ux)=k(x-xs)+xdu;

[0141] Mass transfer coefficient For kinematic viscosity, d h The equivalent diameter of the hollow fiber membrane filament; the molar gas concentration of component 1 within the concentration polarization layer.

[0142] Considering the boundary conditions for concentration polarization:

[0143] When z = 0, x = x0, u = u0, p i =p i,0 p o =p o

[0144] When z = L, v = 0, p o =p o ,

[0145]

[0146] As a specific embodiment, under the condition of obtaining membrane separation, the model is selected according to the condition: the separation process module, the channel model module, and the mass transfer module are selected in sequence, so as to establish and construct the model corresponding to the final simulated experiment.

[0147] Among them, the separation process module is selected as the highest priority, that is, the separation process needs to be determined first, then the logistics channel needs to be determined in the channel model module, and finally the mass transfer model needs to be determined.

[0148] Step S2: Obtain the data parameters and model architecture of the membrane separation to be simulated, and construct the model architecture of the membrane separation to be simulated in the membrane separation model; wherein, the input variables of the data parameters of the membrane separation experiment to be simulated include at least the feed flow rate, the concentration of each component, the permeation rate of each component, the structural parameters of the membrane module, the operating pressure and temperature.

[0149] Step S3: Based on the constructed model architecture, generate simulation data, obtain the model equations under the model architecture, and output the permeate flow and wake flow rates, the molar percentage of each component, and the pressure loss.

[0150] To improve the computational speed of this model, the trapezoidal rule is introduced in this application to perform numerical discretization in the convection-diffusion module and the convection-diffusion module with concentration polarization.

[0151] As a specific example, the model architecture constructed in the concentration-free polarization process includes the following eight differential equations:

[0152] (a1) Equation for conservation of total mass: du=dv;

[0153] (a2) Material conservation equation for component 1: udx + xdu = vdy + ydv;

[0154] (a3) Material conservation equation for component 2: -udx + (1-x)dv = -vdy + (1-y)dv;

[0155] (a4) Permeability equation for component 1: J A (p i xp o y)dz / L=-udx-xdu;

[0156] (a5) Permeability equation for component 2: J A [p i (1-x)-p o (1-y)]dz / (Lα)=udx-(1-x)du;

[0157] (a6) Pressure equation:

[0158] (a7) Mixture viscosity: μ i Indicates the viscosity of component i;

[0159] (a8) Boundary conditions are set as follows:

[0160] At Z=0, x=x0, u=u0, p i =p i,0 p o =p o

[0161] At Z = L, v = 0, p o =p o ,

[0162] In the above series of equations, based on the method described above, the trapezoidal rule is introduced here, and the specific optimization steps are as follows:

[0163] Step A1: By introducing the equation du = dv, the change in flow rate on the permeable side is correlated with the change in flow rate on the residual side, that is, the decrease in flow rate on the permeable side is equal to the increase in flow rate on the residual side.

[0164] Step A2: Substitute equation (a1) into equations (a2) and (a3) ​​to obtain the updated mass balance equation:

[0165] udx + xdv = vdy + ydv……(a2)

[0166] -udx+(1-x)dv=-vdy+(1-y)dv……(a3)

[0167] Step A3: Solve using the updated mass balance equation

[0168] Step A4: Substitute the dv expression into the permeability equation (a4) to obtain the updated permeability rate equation:

[0169] Further simplification yields:

[0170] Step A5: Rearrange equation (a5) to obtain the expression for dx:

[0171] Further simplification yields:

[0172]

[0173] Step A6: Solve the pressure drop equation by combining equations (a6) and (a7).

[0174] Step A7: Substitute the boundary conditions into equation (a8) to obtain the solutions for each output variable.

[0175] As a specific example, some equations and boundary conditions differ for the concentration polarization model compared to the non-concentration polarization model.

[0176] Equation (b1): du=dv

[0177] Equation (b2): udx + xdu = vdy + ydv

[0178] Equation (b3): ​​-udx + (1-x)dv = -vdy + (1-y)dv

[0179] Equation (b4): J A (p i x s -p o y)dz / L=-udx-xdu

[0180] Equation (b5): J A [p i (1-x s )-p o (1-y)]dz / (Lα)=udx-(1-x)du

[0181] Equation (b6): d(ux)=k(xx)s )+xdu

[0182] Equation (b7):

[0183] Equation (b8):

[0184] Equation (b9):

[0185] Boundary conditions:

[0186] At Z=0, x=x0, u=u0, p i =p i,0 p o =p o

[0187] At Z = L, v = 0, p o =p o ,

[0188] In the above series of equations, based on the method described above, the trapezoidal rule is introduced here, and the specific optimization steps are as follows:

[0189] Step B1: Solve equation (b8):

[0190] Integrate over the interval z:

[0191] Assuming μ, u, N, d i For constants:

[0192] Substituting the pressure boundary condition, p i (0)=p i，0 p i (L)=p i，L :

[0193]

[0194] therefore, Here p i,0 p is the feed pressure at the position where the membrane tube length is 0. i，L The feed pressure is at a position where the membrane tube length is L;

[0195] Step B2: Solve equation (b6): d(ux)=k(xx) s )+xdu

[0196] Let equation (b8)x s Substituting the expression into the equation and integrating it over the interval z, we obtain:

[0197] Considering the complexity of the equation, direct integration is difficult to solve. A similar solution is performed according to steps A1 to A5 in the invention description.

[0198] Step B3: Substitute the boundary conditions and solve equations (b1) to (b5) one by one:

[0199] For the differential equation du=dv, we have u=v+C1

[0200] For equation (b2), we have

[0201] Solving equation (b3) yields:

[0202] For equation (b4), we have

[0203] For equation (b5), we have

[0204] Step B4: Substitute the boundary conditions into the above equations (b1) to (b5).

[0205] Substituting the boundary conditions at z = 0 and z = L into equations (b1) to (b5) above, we can obtain x(z), y(z), and p. i (z), p o The analytical solutions of (z), u(z), and v(z) are the concentration distribution, pressure distribution, and velocity distribution along the membrane channel.

[0206] Since the above equations (b1) to (b5) are relatively complex, their rate of change can be set as a constant or solved using numerical methods.

[0207] As a specific embodiment, this application can introduce a numerical solution method for the above equations:

[0208] This invention employs the trapezoidal rule algorithm to solve the integral equation:

[0209]

[0210] By dividing the integration interval into n subintervals, the original integration function is approximated infinitely:

[0211]

[0212] Where x0 = a,x n =b.

[0213] Taking equation (b8) as an example, this invention illustrates how to achieve numerical integration using the trapezoidal rule:

[0214] Discretize it into n subintervals on [0, L]: z0 = 0, z1 = Δz, z2 = 2Δz, ..., z n =L, where Δz = L / n. Then equation (b8) can be transformed into the following form, and thus the solution can be obtained:

[0215]

[0216] For equation (11): d(ux)=k(xx) s )+xdu,

[0217] in

[0218] The numerical integration of equation (b6) using the trapezoidal rule includes the following steps:

[0219] Step C1: Discretize the integration interval [0, L] into n subintervals: z0 = 0, z1 = Δz, z2 = 2Δz, ..., z n =L, where Δz = L / n.

[0220] Step C2: Establish the discrete algebraic equation for estimating the integral function:

[0221]

[0222] Step C3: Apply the discrete algebraic equation to equation (b6).

[0223]

[0224] definition:

[0225] Applying the trapezoidal rule to the function f(z):

[0226]

[0227] Step C4: Discretize x(z) and x s (z): In every z i Calculate x(z) at a point i ) and x s (z i ).

[0228] Step C5: Iterative solution. Based on each x(z) i ) and x s (z i The value of ux(z) is iterated and updated. i ).

[0229] As a specific example:

[0230] Suppose that the interval [0, L] is divided into 4 subintervals, and solve equation (b6).

[0231] Step C1: z0 = 0, z1 = L / 4, z2 = L / 2, z3 = 3L / 4, z4 = L.

[0232] Step C2: Calculate x(z) i ) and x s (z i )value.

[0233] Step C3: According to Calculate f(z) i The value of ).

[0234] Step C4: Perform an equivalent permutation of the integral function f(x) on the interval [0, L].

[0235]

[0236] Step C5: In each z i Iterative update of ux(z) i ).

[0237] Furthermore, to address the potential accumulation of computational errors caused by nonlinear iteration during the discrete computation process of this model, an error module is constructed. This module fits the computational results of the model with the actual values ​​using a polynomial equation. Based on existing experimental data, a multinomial correlation equation is obtained through mapping regression fitting between the model's computational results and the actual values. This multinomial correlation equation is then used to further refine each output result, thereby achieving higher precision training.

[0238] As a specific implementation, the prediction results and corresponding real data from previous predictions can be obtained, and the error between the two can be fitted using a polynomial to obtain an error correction value. Train the output neural network model, where All are matrices, and the polynomial fitting is adjusted to obtain the final prediction correction model.

[0239] The accuracy of the prediction method in this application is explained below with reference to experimental data:

[0240] Scenario 1: A separation experiment of a carbon dioxide (CO2) and nitrogen (N2) mixture was conducted using a tubular membrane with a length of 100 mm, an outer diameter of 12 mm, and an inner diameter of 8 mm, operating at 35℃. The tubular membrane module had an inner diameter of 20 mm, with a single ceramic membrane and a single tubular membrane installed within each module. The mixed gas flowed through the shell side (20 mm inner diameter), and the separated CO2 was collected from inside the membrane tube. The gas separation process employed a counter-current flow. The main focus was on investigating the effect of different feed pressures on the separation performance. Specific experimental data and calculation results are shown in Tables 1-1 to 1-5 below.

[0241]

[0242] Table 1-1 Comparison of Selective Data

[0243]

[0244] Table 1-2 Comparison of Permeate Flow Data

[0245]

[0246] Table 1-3 Comparison of Osmosis Concentration Data

[0247]

[0248] Table 1-4 Comparison of Infiltration Flow Data

[0249]

[0250] Table 1-5 Comparison of Osmotic Concentration Data

[0251]

[0252] Scenario 2: Under operating conditions of 35℃ and feed pressure of 300kPa, a tubular membrane with a length of 100mm, an outer diameter of 12mm, and an inner diameter of 8mm was used to separate a mixture of carbon dioxide (CO2) and nitrogen (N2). The tubular membrane module had an inner diameter of 20mm, and a single tubular membrane was installed within the module. The mixed gas flowed through the shell side, and the separated CO2 was collected from inside the membrane tube. The gas separation process employed a counter-current flow. The main focus was on investigating the effect of different feed concentrations (compositions) on the separation performance. Specific experimental data and corresponding calculation results from the tested software are shown in Tables 2-1 to 2-5 below.

[0253]

[0254] Table 2-1 Comparison of Selective Data

[0255]

[0256] Table 2-2 Comparison of Infiltration Flow Data

[0257]

[0258] Table 2-3 Comparison of permeate concentration data

[0259]

[0260] Table 2-4 Comparison of Infiltration Flow Data

[0261]

[0262] Table 2-5 Comparison of Osmotic Concentration Data

[0263]

[0264] Scenario 3: A separation experiment of a carbon dioxide (CO2) and nitrogen (N2) mixture was conducted using a tubular membrane with a length of 100 mm, an outer diameter of 12 mm, and an inner diameter of 8 mm, under a feed pressure of 300 kPa. The tubular membrane module had an inner diameter of 20 mm, and a single tubular membrane was installed within the module. The mixed gas flowed through the shell side, and the separated CO2 was collected from inside the membrane tube. The gas separation process employed a counter-current flow. The main focus was on investigating the effect of different operating temperatures on the separation performance. Specific experimental data and calculation results from the tested software are shown in Tables 3-1 to 3-5 below.

[0265]

[0266] Table 3-1 Comparison of Selective Data

[0267]

[0268] Table 3-2 Comparison of Permeate Flow Data

[0269]

[0270] Table 3-3 Comparison of Permeate Concentration Data

[0271]

[0272] Table 3-4 Comparison of Infiltration Flow Data

[0273]

[0274] Table 3-5 Comparison of Osmotic Concentration Data

[0275]

[0276] Scenario 4: A zinc chloride (ZnCl2) solution filtration experiment was conducted using a hollow fiber membrane with a length of 100 mm, an outer diameter of 1 mm, and an inner diameter of 0.8 mm, operating at 25℃. The tubular membrane module had an inner diameter of 20 mm and contained 10 hollow fiber membranes. The feed solution flowed through the shell side, while the permeate was collected from inside the fiber membrane tubes. The separation process employed a counter-current flow. The main focus was on investigating the effects of different feed flow rates and feed concentrations (compositions) on the membrane separation performance. Specific experimental data and calculation results from the tested software are shown in Tables 4-1 to 4-6 below.

[0277]

[0278] Table 4-1 Comparison of Retention Rate Data

[0279]

[0280] Table 4-2 Comparison of Infiltration Flow Data

[0281]

[0282]

[0283] Table 4-3 Comparison of Permeate Concentration Data

[0284]

[0285] Table 4-4 Comparison of Infiltration Flow Data

[0286]

[0287] Table 4-5 Comparison of Osmotic Concentration Data

[0288]

[0289] Table 4-6 Comparison of Shell-Side Pressure Drop Data

[0290]

[0291] Scenario 5: A sodium sulfate (Na2SO4) solution filtration experiment was conducted using a hollow fiber membrane with a length of 100 mm, an outer diameter of 1 mm, and an inner diameter of 0.8 mm, operating at 25℃. The tubular membrane module had an inner diameter of 20 mm and contained 10 hollow fiber membranes. The feed solution flowed through the shell side, while the permeate was collected from inside the fiber membrane tubes. The separation process employed a countercurrent flow. The main focus was on investigating the effects of different feed pressures and feed concentrations (compositions) on the membrane separation performance. Specific experimental data and calculation results from the testing software are shown in Tables 5-1 to 5-6 below.

[0292]

[0293] Table 5-1 Comparison of Retention Rate Data

[0294]

[0295] Table 5-2 Comparison of Permeate Flow Data

[0296]

[0297] Table 5-3 Comparison of Permeate Concentration Data

[0298]

[0299] Table 5-4 Comparison of Infiltration Flow Data

[0300]

[0301] Table 5-5 Comparison of Osmotic Concentration Data

[0302]

[0303]

[0304] Table 5-6 Comparison of Shell-Side Pressure Drop Data

[0305]

[0306] Scenario 6: A sodium sulfate (Na₂SO₄) solution filtration experiment was conducted using a hollow fiber membrane with a length of 100 mm, an outer diameter of 1 mm, and an inner diameter of 0.8 mm, under a transmembrane pressure of 500 kPa. The tubular membrane module had an inner diameter of 20 mm and contained 10 hollow fiber membranes. The feed solution flowed through the shell side, and the permeate was collected from inside the fiber membrane tubes. The separation process employed a countercurrent flow. The main focus was on investigating the effect of different operating temperatures on the membrane separation performance. Specific experimental data and calculation results from the tested software are shown in Tables 6-1 to 6-6 below.

[0307]

[0308] Table 6-1 Comparison of Retention Rate Data

[0309]

[0310] Table 6-2 Comparison of Permeate Flow Data

[0311]

[0312]

[0313] Table 6-3 Comparison of Permeate Concentration Data

[0314]

[0315] Table 6-4 Comparison of Infiltration Flow Data

[0316]

[0317] Table 6-5 Comparison of Osmotic Concentration Data

[0318]

[0319] Table 6-6 Comparison of Shell-Side Pressure Drop Data

[0320]

[0321] Scenario 7: Under operating conditions of 35℃ and feed pressure of 400kPa, a tubular membrane with a length of 100mm, an outer diameter of 12mm, and an inner diameter of 8mm was used to separate a helium-methane mixture. The tubular membrane module had an inner diameter of 20mm, and a single ceramic membrane was installed inside the module. The mixed gas flowed through the shell side, and the separated helium was collected from inside the membrane tube. The gas separation process adopted a countercurrent flow.

[0322] The effects of different feed flow rates and feed concentrations (compositions) on separation performance were mainly investigated. Specific experimental data and calculation results from this application are shown in Tables 7-1 to 7-5 below:

[0323]

[0324]

[0325] Table 7-1 Comparison of Selective Data

[0326]

[0327] Table 7-2 Comparison of Permeate Flow Data

[0328]

[0329] Table 7-3 Comparison of Permeate Concentration Data

[0330]

[0331] Table 7-4 Comparison of Infiltration Flow Data

[0332]

[0333] Table 7-5 Comparison of Osmotic Concentration Data

[0334]

[0335] Scenario 8: A separation experiment of a helium (He) and methane (CH4) mixture was conducted using a tubular membrane with a length of 100 mm, an outer diameter of 12 mm, and an inner diameter of 8 mm, under a feed pressure of 300 kPa. The tubular membrane module had an inner diameter of 20 mm, and a single ceramic membrane was installed inside the module. The mixed gas flowed through the shell side, and the separated helium (He) was collected from inside the membrane tube. The gas separation process adopted a countercurrent flow. The main focus was on investigating the effect of different operating temperatures on the separation performance. Specific experimental data and calculation results from this application are shown in Tables 8-1 to 8-5 below.

[0336]

[0337] Table 8-1 Comparison of Selective Data

[0338]

[0339] Table 8-2 Comparison of Permeate Flow Data

[0340]

[0341]

[0342] Table 8-3 Comparison of Permeate Concentration Data

[0343]

[0344] Table 8-4 Comparison of Infiltration Flow Data

[0345]

[0346] Table 8-5 Comparison of Osmotic Concentration Data

[0347]

[0348] Scenario 9: At an operating temperature of 25℃, a tubular membrane with a length of 100mm, an outer diameter of 12mm, and an inner diameter of 8mm was used to conduct retention experiments on dextran solutions of different molecular weights. The tubular membrane module had an inner diameter of 20mm, and a single ceramic membrane was installed inside the module. The filtered liquid was collected from inside the membrane tube, and cross-flow filtration was used for liquid separation. The effects of different feed pressures and glucose molecular weights on retention performance were mainly investigated. Specific experimental data and calculation results from this application are shown in Tables 9-1 to 9-6 below.

[0349]

[0350]

[0351] Table 9-1 Comparison of Retention Rate Data

[0352]

[0353] Table 9-2 Comparison of Permeate Flow Data

[0354]

[0355] Table 9-3 Comparison of Permeate Concentration Data

[0356]

[0357] Table 9-4 Comparison of Infiltration Flow Data

[0358]

[0359]

[0360] Table 9-5 Comparison of Osmotic Concentration Data

[0361]

[0362] Table 9-6 Comparison of Shell-Side Pressure Drop Data

[0363]

[0364] Based on the above nine experiments, we can see that:

[0365] Under simulated membrane separation experimental parameters in six different scenarios, membrane separation process data for carbon dioxide and nitrogen mixtures and sodium sulfate and aqueous solutions were calculated. The six scenarios were: membrane separation of carbon dioxide and nitrogen mixtures under different feed pressures; membrane separation of carbon dioxide and nitrogen mixtures under different feed concentrations (compositions); membrane separation of carbon dioxide and nitrogen mixtures under different operating temperatures; membrane separation of zinc chloride and aqueous solutions under different feed flow rates and feed concentrations (compositions); membrane separation of sodium sulfate and aqueous solutions under different feed pressures and feed concentrations (compositions); and membrane separation of sodium sulfate and aqueous solutions under different operating temperatures. The calculated data were compared with experimental data under the same conditions. The relative deviations of the selectivity, permeate flow rate, permeate concentration, residual flow rate, residual concentration, rejection rate, and shell-side pressure drop data in the six scenarios were all less than 20%, indicating good agreement between the two sets of data.

[0366] Under simulated membrane separation experimental parameters in three different scenarios, membrane separation process data for helium-methane mixtures, dextran, and aqueous solutions were calculated. The three membrane separation experiments included: membrane separation of helium-methane mixtures under different feed flow rates and feed concentrations (compositions); membrane separation of helium-methane mixtures under different operating temperatures; and membrane separation of dextran and aqueous solutions under different feed pressures and glucose molecular weights. The calculated data were compared with experimental data under the same conditions. The relative deviations of the selectivity, permeate flow rate, permeate concentration, residual flow rate, residual concentration, rejection rate, and shell-side pressure drop data in the three experimental scenarios were all less than 20%, indicating good agreement between the two sets of data.

[0367] After compiling the solution of the present invention into software, based on actual calculations, the calculation time for the membrane separation process of the oxygen and nitrogen mixture was 19.56 milliseconds, 18.72 milliseconds, and 17.83 milliseconds for three executions, with an average calculation time of 18.70 milliseconds.

[0368] The calculations for the membrane separation process of a nitrogen and hydrogen mixture were performed, and the calculation times for three executions were 21.36 ms, 20.52 ms, and 19.61 ms, respectively, with an average calculation time of 20.50 ms.

[0369] It can basically compress each simulation calculation process to about 20 milliseconds, and the error between the final data parameters and the experimental data does not exceed 20%, which is very reliable.

[0370] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0371] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations (such as quantity, shape, position, etc.) can be made to the technical solutions of the present invention, and these equivalent transformations are all protected by the present invention.

Claims

1. A method for predicting the separation performance of selective permeation membranes, characterized in that, Includes the following steps: Step S1: Construct a membrane separation model based on the dynamic relationship between the feed flow and the permeate flow in membrane separation; wherein, the membrane separation model includes a separation process module, a channel model module, and a mass transfer module; the separation process module is constructed based on the flow direction of the feed flow and the permeate flow; the channel model module is constructed by selecting the corresponding flow channel based on the direction of the feed flow; the mass transfer module is constructed based on the gas transfer relationship across the membrane, the change relationship along the length of the membrane module, and whether concentration polarization occurs in the membrane separation model; Step S2: Obtain the data parameters and model architecture of the membrane separation to be simulated, and construct the model architecture of the currently simulated membrane separation in the membrane separation model; wherein, the input variables of the data parameters of the membrane separation to be simulated include at least the feed flow rate, the concentration of each component, the permeation rate of each component, the structural parameters of the membrane module, the operating pressure and temperature; Step S3: Based on the constructed model architecture, generate simulation data, obtain the model equations under the model architecture, and correlate the flow rate changes on the permeate side and the flow rate changes on the residual side with the model equations to quickly obtain the changes in permeate flow rate, residual flow rate, concentration of each component, and pressure of each component. The mass transfer module includes: an interface transfer module and a convection diffusion module; The interface transfer module is based on a model of mass transfer across the membrane, and constructs a permeation rate model for each component based on the relationship between feed flow rate, permeate flow rate, and residual flow rate. When the difference in permeation rates of each component does not exceed a preset value, no concentration polarization phenomenon occurs in the membrane separation model. Then, the convection diffusion module constructs the permeation model and boundary conditions for each component based on the relationship between the concentration gradient, pressure gradient and velocity gradient along the length of the membrane module. When the difference in permeation rates among the components exceeds a preset value, concentration polarization occurs in the membrane separation model. Then, the convection diffusion module introduces a concentration polarization model based on the concentration gradient, pressure gradient, velocity gradient and concentration polarization phenomenon along the length of the membrane module to construct the permeation model of each component, as well as the boundary conditions, mass transfer driving force of the concentration polarization layer and the molar gas concentration of each component in the concentration polarization layer. When concentration polarization occurs in the membrane separation model, the corresponding interface transfer equation is: ,in, ; in, This indicates the fluid flow rate on the permeate side of the membrane module. This indicates the molar concentration on the permeate side of the membrane module. This indicates the molar concentration on the permeate side of the membrane module. Indicates the external pressure of the silk. J1 represents the internal pressure of the fiber, J2 represents the permeability coefficient, and J2 represents the permeability coefficient of the difficult-to-permeate component. The mass transfer coefficient; The interface transfer equations are optimized and solved, including: Step 1: Integral interval Discretized Sub-intervals: ,in ; Step 2: Establish an alternative equation for the integral equation based on the trapezoidal rule: ； Step 3: Apply the substitution equation to solve the interface transfer equation: ； definition: , Establish an alternative equation for solving the interface transfer equation based on the trapezoidal rule: ; Step 4: Calculation : Calculate each and And thus obtain ; Step 5: Iterative solution, based on each Value, iterate and update distance Permeation flux at point .

2. The method for predicting the selective permeation membrane separation performance according to claim 1, characterized in that, Step S3 involves relating the flow rate changes on the permeate side and the residual flow rate on the residual flow side to the model equations to quickly obtain the changes in permeate flow rate, residual flow rate, concentration of each component, and pressure of each component. Specifically, this includes: Based on the trapezoidal rule, the integral interval in the model equation is divided into several trapezoidal infinitesimal elements to obtain the area under the curve. By constructing an operator in each infinitesimal element in which the decrease in permeable flow is equal to the increase in residual flow, and introducing the operator into the model equation under the model framework, the model equation is transformed. Under the constraint of boundary conditions, the changes in permeable flow, residual flow, concentration of each component, and pressure of each component are quickly obtained.

3. The method for predicting the selective permeation membrane separation performance according to claim 2, characterized in that, The method of dividing the integration interval in the model equation into several trapezoidal infinitesimal elements based on the trapezoidal rule to obtain the area under the curve specifically includes: Based on the trapezoidal rule, the model equations In the integration interval Transform the above: ； based on Transformation of the model equations for an equidistant trapezoidal infinitesimal element: ,in, .

4. The method for predicting the selective permeation membrane separation performance according to claim 1, characterized in that, The separation process module includes: a counter-current unit, a parallel-current unit, and a cross-current unit; In the countercurrent unit, the flow directions of the feed stream and the permeate stream on both sides of the membrane are opposite; In the parallel flow unit, the feed stream and the permeate stream on both sides of the membrane flow in the same direction; In the cross-flow unit, the flow directions of the feed stream and the permeate stream on both sides of the membrane are intersected. In the approximation, it is assumed that the two flow directions are the same and that there is a difference in concentration boundaries.

5. The method for predicting the selective permeation membrane separation performance according to claim 1, characterized in that, The channel model module includes: tube side units and shell side units; In the tube-side unit, easily permeable substances in the feed stream permeate from inside the membrane fibers to the shell side of the membrane separation device; In the shell-side unit, easily permeable substances in the feed stream permeate from the outside of the membrane fibers to the inside of the membrane fibers; In the channel model module, a total mass conservation model, a material conservation model, and a pressure model are constructed.

6. The method for predicting the selective permeation membrane separation performance according to claim 1, characterized in that, The selective permeation membrane separation performance prediction method further includes: Step S4: Obtain the prediction results and corresponding real data from each prediction, perform polynomial fitting on the error between the two to obtain the error correction value, and then... Train the output neural network model, where and All are matrices, and the polynomial fitting is adjusted to obtain the final prediction correction model.

7. The method for predicting the selective permeation membrane separation performance according to claim 1, characterized in that, The variation of the membrane separation model in the length direction is calculated using the following equation: Equation for calculating residual flow rate u: , The flow rate at node i on the permeate side; Equation for calculating permeation flow rate v: , Let i be the flow rate at node i on the permeation side; Equation for calculating residual concentration x: , The concentration of the component on the osmotic side at the i-th node; Equation for calculating osmotic concentration y: , Let be the component concentration on the permeation side at node i. ‌ 8. The method for predicting the selective permeation membrane separation performance according to claim 1, characterized in that, The membrane separation model has at least two components; The variation of any component j is calculated using the following model: Permeate flow Calculation equation: ; Infiltration flow Calculation equation: ; residual concentration Calculation equation: ; Osmotic Concentration Calculation equation: ; The total pressure is the sum of the pressures of each component: m is the number of components in the system; The total residual flow rate u is the sum of the residual flow rates of each component: u ; The total permeation flow rate v is the sum of the permeation flow rates of each component: v ; The sum of the residual concentrations calculated as percentages is 1: =1; The sum of the osmotic concentrations calculated as percentages is 1: =1.

Images