A method for optimizing a decarbonated seawater feed reverse osmosis desalinated boron-free seawater desalination system considering uncertainty

Abstract

The application discloses a kind of decarbonization seawater feed reverse osmosis boron removal seawater desalination system optimization method considering uncertainty.The method establishes the numerical model of reverse osmosis boron removal transfer mechanism in a strict mechanism way, the scaling tendency of carbonic acid in seawater is greatly reduced after removal, the pH value of each stage reverse osmosis is increased, the boron retention rate is greatly improved, the superstructure model of reverse osmosis system is established, the changes of brine pressure, concentration, flow and pH value along the pressure vessel axis are considered, the material distributor and mixer consider the balance of material, salinity, boron concentration and pH value, and the operation condition constraint is added to ensure the safe operation of system.The two-stage stochastic optimization model is used to consider the uncertainty factors in the operation process, so that the produced water meets the standard of boron content in drinking water, the process and operation condition are adjusted according to different scenarios to further reduce energy consumption, which is more robust, provides theoretical basis and technical reference for industrial application, and has very good application prospect.

Description

Technical Field

[0001] The present invention belongs to the field of seawater treatment, and specifically relates to optimizing a reverse osmosis deboronation system for decarbonized seawater feed, taking into account uncertainties in the operation process, and effectively reducing system energy consumption while ensuring the deboronation rate. Background Art

[0002] Reverse osmosis desalination, due to its mature technology and continuously decreasing energy consumption, has dominated the international and domestic desalination markets. Natural seawater contains approximately 4-6 mg / L of boron, primarily in the form of H₃BO₃ boric acid molecules, which easily permeate membranes at a typical pH of 7.9-8.2. Reverse osmosis membranes are less than ideal for removing boron from seawater. Despite the continuous development of advanced RO membranes, the removal rate in commercial systems is only 78%-80% or even lower. Consequently, boron levels in the resulting water exceed 1.0 mg / L. Excessive boron levels can cause reproductive and neurological problems in humans. my country's current "Standard for Drinking Water Quality (GB5479-2006)" stipulates that the boron content in drinking water should be less than 0.5 mg / L. Excessive boron levels can also harm certain crops. my country's agricultural irrigation water quality standards require a boron tolerance of 1 mg / L for cucumbers, beans, potatoes, winter squash, leeks, and onions, while lemons and blackberries tolerate only 0.5 mg / L. Traditional reverse osmosis designs cannot meet these boron requirements for drinking water and other industrial and agricultural water.

[0003] To reduce water production costs and energy consumption, numerous reports have been published on the optimization of reverse osmosis (RO) desalination systems, including membrane element placement optimization within pressure vessels, network-based superstructure optimization, global optimization, and multi-objective optimization. Optimizing the design of RO desalination technology to improve energy efficiency is a key approach to reducing desalination costs. While energy optimization of ideal RO systems under thermodynamic constraints can provide theoretical guidance for system design, many factors must be considered in practical applications. Some researchers have implemented 24-hour and annual operation scheduling for single-stage RO systems by adjusting the frequency of variable-frequency high-pressure pumps, membrane module on / off cycles, and operating conditions. Using LabVIEW, which leverages optimization and model predictive control, can effectively reduce system energy consumption. Many domestic research institutions, such as Tianjin University, Ocean University of China, and the State Oceanic Administration's Water Treatment Center, have conducted research on desalination system simulation and optimization, achieving significant results in full-process spiral-wound RO desalination system operation scheduling, membrane element cleaning and replacement optimization, model predictive control, robust system scheduling under uncertainty, and optimization of heat-membrane-coupled cogeneration systems. These studies help further reduce system energy consumption and water production costs, but they fail to consider the impact of boron removal on system design and operational scheduling. In recent years, there have been some reports on the optimization of reverse osmosis deboronization systems, examining the impact of factors such as temperature, pH, and boron content requirements on reverse osmosis system design. However, no research has yet considered the impact of uncertainty in optimizing reverse osmosis deboronization systems using decarbonized seawater feed. Summary of the Invention

[0004] The present invention discloses a method for optimizing a decarbonized seawater feed reverse osmosis deboronation seawater desalination system taking into account uncertainty. The method adopts a strict mechanism to establish an accurate model of the reverse osmosis transfer mechanism taking into account deboronation, and is described by differential and algebraic equations. After the carbonic acid in the seawater is removed, the scaling tendency is greatly reduced, and each level of reverse osmosis can increase the pH value of the inlet water, and the retention rate is greatly improved. A superstructure model of the reverse osmosis system is established, taking into account the changes in the brine pressure, concentration, flow rate and pH value along the axial direction of the pressure vessel, as well as the increase in salinity due to brine mixing in the power exchanger. The material distributor and mixer take into account the balance of materials, salinity, boron concentration and pH value, and add operating condition constraints to ensure the safe operation of the system. A two-level stochastic optimization model is used to consider the uncertainty factors in the operation process, and the process and operating conditions are adjusted according to different scenarios to further reduce the energy consumption of the system, which is more robust. The present invention comprehensively considers the impact of uncertain factors on the seawater desalination system, and strives to further reduce the energy consumption of the seawater desalination system so that the water produced by the system meets the boron content standard for drinking water. This is of great significance for energy conservation and emission reduction, provides a theoretical basis and technical reference for its industrial application, and has very good application prospects.

[0005] The present invention comprises the following steps:

[0006] 1. A method for optimizing a decarbonized seawater feed reverse osmosis deboronized seawater desalination system taking into account uncertainty, characterized in that the method comprises the following steps:

[0007] Step 1: Establish a spiral membrane element reverse osmosis desalination process model;

[0008] According to the reverse osmosis process mechanism, differential equations are used to describe the axial changes of salinity, boron concentration, pressure, and flow in the pressure vessel. The finite difference method differential equation is discretized to obtain the following equation:

[0009]

[0010] Among them, A, B, B TB are the permeability coefficients of pure water, salt and boron respectively, P represents pressure, C represents salinity, π represents osmotic pressure, ρ p and V w represent the density and flow rate of fresh water, J w and J s are pure water flux and salt permeation flux, V w is the permeation flow rate, T is the temperature, K is the mass transfer coefficient, d e is the equivalent diameter of the feed channel, S l is the area of ​​a differential unit of the membrane element, S l =S m ·n m / L,S m is the area of ​​a single membrane element, n m is the number of pressure vessel membrane elements, L is the total number of differential unit nodes, Re is the Reynolds number, Re=ρVd e / μ, where μ is the dynamic viscosity, Sc is the Schmidt number, Sc=μ / ρD s , D s is the diffusion coefficient of salt, Q is the flow rate, V is the feed flow rate, V = Q / (3600S fcs ε sp ), S fcs is the cross-sectional area of ​​the feed channel, ε sp is the porosity of the feed channel screen, σ is the reflection coefficient, α0 and α1 are the fractions of boric acid and borate ions, respectively, and pK a is the primary ionization constant of boric acid, K λ is the friction coefficient, FF d is the contamination coefficient, e is the activation energy of the membrane (when T≤298K, e=25,000J / mol -1 , when T>298K, e=22,000J / mol -1 ), R is the gas constant, B inis the annual salt permeation increase rate, N mlp is the average life of the reverse osmosis membrane, N RO Represents the total reverse osmosis level, k λ represents the internal friction coefficient of the membrane element, the subscript ch is the feed or product water flow channel of the membrane element, b is the concentrated brine, boric is the boric acid molecule, borate is the borate ion, f is the feed seawater, p is the product water, mw is the membrane surface, ref is the membrane parameter when T0 is free of contamination at 298K, l is the differential unit node, and j represents the jth reverse osmosis stage;

[0011] Boundary conditions for the finite difference method:

[0012] z=0,V=V in , Q=Q in , C boric =C boric，in , C berate =C borate，in , C=C in , P=P in ;

[0013] Where in represents the pressure vessel inlet;

[0014] The water production diversion design directly delivers the water produced at the front end of the pressure vessel to the final water production, and the water produced at the back end enters the next stage of reverse osmosis for desalination. Flow regulating valves are added at both ends of the pressure vessel to adjust the water output ratio at both ends. The model is expressed as follows:

[0015] Q ch，b，l+1 =Q ch，b，l -3600V w，l S l (15)

[0016] Q ch，b，l+1 C ch，b，l+1 -Q ch，b，l C ch，b，l =-3600V w，l S l C ch，p，l (16)

[0017] Q ch，b，l+1 C TB，ch，b，l+1 -Q ch，b，l C TB，ch，b，l =-3600V w，l S l C TB，ch，o，l (17)

[0018]

[0019] Q f，n =Q b，n +Q p，n，lc +Qp，n，hc (twenty four)

[0020] Q f，n C f，n =Q b，n C b，n +Q p，n，lc C p，n，lc +Q p，u，hc C p，n，hc (25)

[0021] Q f，n C TB，f，n =Q b，n C TB，b，n +Q p，n，lc C TB，p，n，lc +Q p，n，hc C TB，p，n，hc (26)

[0022] Y l -Y l+1 ≥0 (27)

[0023] Wherein the binary variable Y represents the water flow direction of the differential unit in the pressure vessel, the subscript lc represents the water produced at the front end of the pressure vessel, hc represents the water produced at the rear end of the pressure vessel, and n represents the nth pressure vessel. Formula (27) indicates that the water produced at the front end or rear end of each differential unit in the pressure vessel has a consistent flow direction.

[0024] Salt water osmotic pressure π, dynamic viscosity μ and salt diffusion coefficient D s Calculated by the following fitting formula:

[0025] π=4.54047(10 3 C / M s ρ) 0.987 (28)

[0026] μ=(1.4757×10 -3 +2.4817×10 -6 C+9.3287×10 -9 C 2 )exp(-0.02008T) (29)

[0027] D s =6.725×10 -6 exp(0.1546×10 -3 C-2513 / (T+273.15)) (30)

[0028] Among them, M s is the molar mass of the solute;

[0029] Step 2. Establish a reverse osmosis superstructure model;

[0030] The reverse osmosis system includes seawater intake and pretreatment, water post-treatment, reverse osmosis membrane group, pump, pressure exchanger (PX), logistics mixer and separator. Due to the presence of carbonate ions in seawater, precipitation is easily formed during the reverse osmosis operation. The seawater is acidified and the pH value is adjusted to 4.0. The free CO2 in the seawater is removed by a fan. Decarbonized seawater can effectively reduce the tendency of scaling. Adding strong alkali to raise the pH value to alkaline effectively increases the boron retention rate. The reverse osmosis superstructure contains N PS boost stages and N RO There are N reverse osmosis stages in total. PS +2 logistics nodes, 2 refers to the salt water and fresh water that eventually leave the reverse osmosis system, N PS Each of the logistics nodes represents a stream that enters a reverse osmosis unit directly after being pressurized by a high-pressure pump (or without being pressurized by a high-pressure pump). Each reverse osmosis stage is composed of multiple parallel pressure vessels, each pressure vessel consists of 2 to 8 membrane elements in series and operates under the same conditions. Each stream of salt water and fresh water leaving the reverse osmosis stage can enter N PS +2 logistics nodes, each logistics is expressed as a function of flow rate, salinity, boron concentration and pressure, each feed M IN After passing through the logistics distributor, it is divided into M OUT Some logistics are combined into one logistics after passing through the logistics mixer. The logistics distributor and mixer are expressed as:

[0031]

[0032] C in，out =C in out=1,...M OUT (32)

[0033] C TB，in，out =C TB，in out=1,...M OUT (33)

[0034] P in，out =P in out==1,...M OUT (34)

[0035] pH in，out =pH in out=1,...M OUT (35)

[0036]

[0037] 0=(P in -P out )Qin，out in=1,...M IN (40)

[0038] Formulas (31)-(35) represent the logistics distributor, formulas (36)-(39) represent the logistics mixer, and formula (40) represents the isobaric mixing constraint, which allows the reverse osmosis stage water to be mixed with the final system water and the reverse osmosis stage high-pressure brine to be depressurized and mixed with the system feed. The total boron concentration C TB Boric acid molecule C boric and borate ion C borate The sum of the concentrations, O in 、C in C TB，in and P in represent the flow rate, salinity, boron concentration and pressure at the inlet of the logistics distributor, respectively, Q out 、C out 、C TB，out and P out represent the flow rate, salinity, boron concentration and pressure at the outlet of the logistics mixer, respectively. Q in，out represents the outlet flow of the logistics distributor, C in，out 、C TB，in，out and P in，out represents the salinity, boron concentration and pressure at the outlet of the logistics distributor, Q acid and Q base are the amounts of acid and base added, respectively, C acid and C base are the concentrations of acid and base, respectively, C H + ，acid and C OH - ，base are the hydrogen ion concentrations in the acid and alkali solutions, respectively, and the subscripts in and out represent the inlet and outlet, respectively;

[0039] The material balance equation of the high-pressure pump and the work exchanger is:

[0040] Q ps，1 =Q hpp +Q pxlin (41)

[0041] Q ps，1 C ps，1 =Q hpp C hpp +Q pxlin C pxlin (42)

[0042] Q ps，1 C TB，ps，1 =Q hpp C TB，hpp +Qpxlin C TB，pxlin (43)

[0043] Q RO，1 =Q hpp +Q pxhout (44)

[0044] Q RO，1 C RO，1 =Q hpp C hpp +Q pxhout C pxhout (45)

[0045] Q RO，1 C TB，RO，1 =Q hpp C TB，hpp +Q pxhout C TB，pxhout (46)

[0046] Q pxhout =Q pxlin (47)

[0047] Q pxhin =Q pxlout (48)

[0048] L px Q pxhin / 100=Q pxhin -Q pxhout (49)

[0049] L px [%]=0.3924+0.01238P pxhin (50)

[0050] C pxhout =Mix(C pxhin -C pxlin )+C pxlin (51)

[0051] C TB，pxhout =Mix(C TB，pxhin -C TB，pxlin )+C TB，pxlin (52)

[0052] Mix=6.0057-0.3559OF+0.0084OF 2 (53)

[0053] OF[%]=100×(Q pxhin ,-Q pxhout ) / Q pxhin (54)

[0054] C pxlout Q pxlout =Q pxlin C pxlin +Q pxhin C pxhin -Q pxhout C pxhout (55)

[0055] C TB，pxlout Q pxlout =Q pxlin C TB，pxlin +Q pxhin C TB，pxhin -Q pxhout C TB，pxhout (56)

[0056] Among them L PX is the leakage rate, Mix is ​​the volume mixing ratio, OF (-10% ≤ OF ≤ 15%) is the lubrication flow rate, the subscripts hpp, pxhin, pxlin, pxhout, and pxhin represent the high-pressure pump, the low-pressure feed seawater and high-pressure brine entering the work exchanger, and the pressurized seawater and decompressed brine leaving the work exchanger, respectively; the subscript ps represents the boosting stage, and RO represents the reverse osmosis stage;

[0057] The material flow leaving the i-th boosting stage directly enters the j-th reverse osmosis stage. The same type of membrane elements are used in the reverse osmosis pressure vessel of the same stage. Its characteristics such as pure water permeability coefficient, solute permeability coefficient, membrane area and feed spacer thickness remain unchanged. The membrane element model k used in the j-th reverse osmosis pressure vessel is determined by the following formula:

[0058]

[0059] Introduce binary variable y j，k When it is 1, it means that the kth type of membrane element is selected in the jth stage of reverse osmosis, otherwise it is 0; Formula (58) defines the maximum allowable water inlet pressure of the membrane element, U is a sufficiently large number, K t It is a collection of types of reverse osmosis membrane elements;

[0060] Formulas (60) and (61) are introduced to determine the brine and product water of the reverse osmosis stage entering the next reverse osmosis stage. If the binary variable β b，i，j =1, the brine of the jth stage enters the next reverse osmosis stage. p，i，j If it is equal to 1, the produced water of the jth stage enters the next reverse osmosis stage. Formula (62) can prevent the concentrated brine and produced water from entering the next reverse osmosis stage at the same time.

[0061]

[0062] As water passes through the membrane, the pH of the brine and product water will also change, which is described by the following formula:

[0063]

[0064] in represents the recovery rate of the j-th stage reverse osmosis, r l，j It represents the recovery rate of each differential unit in the j-th stage reverse osmosis pressure vessel, It can be determined that the feed seawater of the jth reverse osmosis stage is the previous reverse osmosis stage or reverse osmosis section;

[0065] The pH value of seawater is 8.2. To effectively remove CO2 from seawater, use (Q acid，0 +Q acid，1 ) strong acid to adjust the pH value of seawater to 4.0, and then remove the CO2 in the seawater through the purge tower, and then add (Q base，0 +Q base，1 ) to increase the pH value of seawater to increase the boron removal rate.

[0066]

[0067] (Q f +Q acid，0 )10 -7 +Q acid，1 ·C acid =(Q f +Q acid，0 +Q acid，1 )10 -4 (69)

[0068] (Q f +Q acid，0 +Q acld，1 )10 -4 =Q base，0 C base (70)

[0069]

[0070] Wherein formula (71) calculates the amount of strong base required to increase the pH value of the reverse osmosis stage feed;

[0071] The entire reverse osmosis network satisfies the following material balance relationship and product water demand constraints:

[0072] Q f =Q b +Q p (72)

[0073] Q f C f =Qb C b +Q p C p (73)

[0074] Q f C TB，f =Q b C TB，b +Q p C TB，p (74)

[0075]

[0076] Q p ≥Q p，lo (81)

[0077] C p ≤C p，np (82)

[0078] C TB，p ≤C TB，p，up (83)

[0079] Where Q b 、C b and C TB，b are the brine flow rate, salinity and boron concentration leaving the reverse osmosis network, Q p 、C p and C TB，p represent the flow rate, salinity and boron concentration of product water, respectively, and the subscripts lo and up represent the minimum required value and the maximum allowed value, respectively;

[0080] Step 3. System flow and operating condition constraints;

[0081] To ensure the safe operation of the reverse osmosis system, the following constraints are set in the model: the concentration polarization factor is the salt concentration C on the membrane surface. ch.mw.1 and the salinity of the bulk solution C ch，b.1 The concentration polarization factor limit of the first-stage reverse osmosis of decarbonized seawater is 1.22, the concentration polarization factor limit of the first-stage reverse osmosis of traditional seawater feed is 1.2, and the concentration polarization factor of the second-stage reverse osmosis is 1.4 at most because the salt content of its feed water has been significantly reduced; the maximum pressure loss of a single pressure vessel is 0.35 MPa, and the maximum average water production flux of the first and second stages is 20 L / (m 2 h) and 40L / (m 2 ·h), the maximum water flux of the first membrane element of the first stage and the second stage is 35L / (m 2 h) and 48L / (m 2 ·h), the minimum brine flow rate in the first and second stage pressure vessels is 3.6m 3 / h and 2.4m3 / h, the brine concentration is less than 90kg / m 3 , the maximum pH value of reverse osmosis feed is 9.5, and the maximum pH value of reverse osmosis grade feed is 11.0;

[0082] Step 4. Establish a reverse osmosis system optimization design model

[0083] The optimization design problem of the reverse osmosis system is expressed as a mixed integer nonlinear programming, with formula (79) as the objective function and satisfying the process thermodynamics, unit operation, and design requirements constraints:

[0084]

[0085] The objective function Ew takes into account the system energy consumption, total water intake, membrane module scale and acid and alkali reagent addition, ΔP SWIP , ΔP hpp , ΔP bp and ΔP bppx They represent the pressure difference of seawater intake pump, high pressure pump, interstage pump and booster pump respectively, Q f , Q hpp , Q bp and Q bppx They represent the flow rates of feed seawater, high-pressure pump, interstage pump and booster pump respectively, η is the efficiency of the equipment, f c is the load factor of the reverse osmosis device, n m，j and n pvj represents the number of membrane elements in the j-th stage reverse osmosis pressure vessel and the number of the j-th stage reverse osmosis pressure vessel, and the subscript moter is the motor of the pump;

[0086] Step 5. Solve the formed reverse osmosis system optimization problem;

[0087] The optimization objective is Ew formula (84),

[0088] The constraint equations are: reverse osmosis process model formulas (1)-(30),

[0089] Reverse osmosis superstructure model formulas (31)-(83),

[0090] System flow and operating condition constraints;

[0091] A two-level stochastic model is used to optimize the process and operating conditions of the reverse osmosis system under uncertain conditions. The uncertainty parameters in the stochastic optimization model are expressed as probability distribution Ω = {ω1, ω2}, so the two-level stochastic model is expressed as:

[0092]

[0093] constraint:

[0094]

[0095] constraint:

[0096] The first two equations represent the first-level model, and the last two equations represent the second-level model. and are decision variables, including the flow rate Q, pressure P, pH at each level in the reverse osmosis model, and the logistics distributor outlet variable Q in，out , represents the number of membrane elements in the j-th level reverse osmosis pressure vessel n m，j , the number of reverse osmosis pressure vessels in the jth stage n pv，j , binary variable Y l and binary variable y j，k ;c T are the parameters in the objective function, is the mathematical expectation of the optimal solution of the second-level model. Formula (85) represents the constraint equation in the optimization problem of the reverse osmosis system, where are the coefficients of the constraint equations, and is the right side of the constraint equation, T ω and W ω are the transposed and compensation matrices respectively; in the first-level model, n m，j and n pv，j Holding constant, other decision variables are adjusted for uncertainty in the second-level model;

[0097] Mathematical programming software is used to solve the above mixed integer nonlinear programming problem. By assigning different initial values ​​to the variables and iterating from multiple starting points, the optimized process and operating conditions of the system can be obtained.

[0098] Beneficial effects of the present invention:

[0099] The present invention discloses a method for optimizing a decarbonized seawater feed reverse osmosis deboronized seawater desalination system taking uncertainty into consideration. After removing carbonic acid from seawater, the scaling tendency is greatly reduced, and each level of reverse osmosis can increase the pH value of the inlet water, and the retention rate is greatly improved. A secondary stochastic optimization model is used to consider the uncertainty factors in the operation process, and the process and operating conditions are adjusted according to different scenarios to further reduce the energy consumption of the system, and the optimization scheme is more robust. The present invention comprehensively considers the impact of uncertainty factors on the design and operation of the seawater desalination system, and strives to further reduce the energy consumption of the seawater desalination system so that the water produced by the system meets the standard for the boron content in drinking water. It is of great significance to energy conservation and emission reduction, provides a theoretical basis and technical reference for its industrial application, and has very good application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Figure 1 Schematic diagram of the superstructure of the decarbonized feed seawater reverse osmosis system;

[0101] Figure 2 Schematic diagram of the superstructure of a conventional reverse osmosis system;

[0102] Figure 3 Comparison results between the model prediction and experimental values ​​of boron concentration in primary reverse osmosis product water;

[0103] Figure 4 Comparison results between the model prediction and experimental values ​​of pH value of primary reverse osmosis brine;

[0104] Figure 5 Comparison results between the model prediction and experimental values ​​of boron concentration in secondary reverse osmosis product water;

[0105] Figure 6 Comparison results between the model prediction and experimental values ​​of pH value of secondary reverse osmosis brine;

[0106] Figure 7 Comparison results between the model prediction and experimental values ​​of pH value of secondary reverse osmosis brine;

[0107] Figure 8 Optimization process of decarbonized feed seawater reverse osmosis system considering uncertainty of contamination coefficient;

[0108] Figure 9 Optimization process of conventional feed seawater reverse osmosis system considering uncertainty of contamination coefficient;

[0109] Figure 10 Optimization process of decarbonized feed seawater reverse osmosis system considering the uncertainty of contamination coefficient and maximum boron concentration in produced water;

[0110] Figure 11 Optimization process of conventional feed seawater reverse osmosis system considering the uncertainty of contamination coefficient and maximum boron concentration in produced water; DETAILED DESCRIPTION

[0111] The present invention will be further described below with reference to the accompanying drawings:

[0112] The present invention comprises the following steps:

[0113] 1. A method for optimizing a decarbonized seawater feed reverse osmosis deboronized seawater desalination system taking into account uncertainty, characterized in that the method comprises the following steps:

[0114] Step 1: Establish a spiral membrane element reverse osmosis desalination process model;

[0115] According to the reverse osmosis process mechanism, differential equations are used to describe the axial changes of salinity, boron concentration, pressure, and flow in the pressure vessel. The finite difference method differential equation is discretized to obtain the following equation:

[0116]

[0117]

[0118] where A, B, B TB are pure water, salt and boron permeability coefficients, P represents pressure, C represents salinity, π represents osmotic pressure, ρ p and V w are the density and flow rate of fresh water, J w and J s are pure water flux and salt permeation flux, V w is permeation flow rate, T is temperature, K is mass transfer coefficient, d e is equivalent diameter of feed flow channel, S l is the area of a differential element of membrane module, S i = S m · n m / L, S m is the area of a single membrane module, n m is the number of membrane modules in a pressure vessel, L is the total number of differential element nodes, Re is Reynolds number, Re = ρVd e / μ, where μ is dynamic viscosity, Sc is Schmidt number, Sc = μ / ρD s , D s is the diffusion coefficient of salt, Q is flow rate, V is feed flow rate, V = Q / (3600S fcs ε sp ), S fcs is the cross-sectional area of feed flow channel, ε sp is the porosity of feed flow channel screen, σ is reflection coefficient, α0 and α1 are the fractions of boric acid and borate ions, pK a is the first ionization constant of boric acid, K λ is friction coefficient, FF d is fouling coefficient, e is the activation energy of membrane (e = 25,000 J / mol -1 for T ≤ 298 K, e = 22,000 J / mol -1 for T > 298 K), R is the gas constant, B in is the annual increase rate of salt permeation, N mlp is the average life of reverse osmosis membrane, N RO represents the total number of reverse osmosis stages, k λ represents the friction coefficient in membrane module, subscript ch is the feed or product water flow channel of membrane module, b is concentrated brine, boric is boric acid molecule, borate is borate ion, f is feed seawater, p is product water, mw is membrane surface, ref is the parameter of membrane without fouling at 298 K, l is differential element node, j represents the jth reverse osmosis stage.

[0119] Boundary conditions for the finite difference method:

[0120] z=0,V=V in , Q=Q in , C boric =C boric，in , C borate =C borate，in , C=C in , P=P in ;

[0121] Where in represents the pressure vessel inlet;

[0122] The water production diversion design directly delivers the water produced at the front end of the pressure vessel to the final water production, and the water produced at the back end enters the next stage of reverse osmosis for desalination. Flow regulating valves are added at both ends of the pressure vessel to adjust the water output ratio at both ends. The model is expressed as follows:

[0123] Q ch，b，l+1 =Q ch，b，l -3600V w，l S l (15)

[0124] Q ch，b，l+1 C ch，b，l+1 -Q ch，b，l C ch，b，l =-3600V w，l S l C ch，p，l (16)

[0125] Q ch，b，l+1 C TB，ch，b，l+1 -Q ch，b，l C TB，ch，b，l =-3600V w，l S l C TB，ch，p，l (17)

[0126]

[0127]

[0128] Q t，n =Q b，n +Q p，n，lc +Q p，n，hc (twenty four)

[0129] Q f，n C f，n =Q b，n C b，n +Q p，n，lc C p，n，lc +Q p，n，hc Cp，n，hc (25)

[0130] Q f，n C TB，f，n = Q b，n C TB，b，n + Q p，n，lc C TB，p，n，lc + Q p，n，hc C TB，p，n，hc (26)

[0131] Y l - Y l+1 ≥ 0 (27)

[0132] where binary variable Y represents the flow direction of water production in the differential unit of pressure vessel, subscript lc represents water production at the front end of pressure vessel, hc represents water production at the rear end of pressure vessel, and n represents the nth pressure vessel; formula (27) represents that water production in each differential unit at the front end or the rear end of pressure vessel has a consistent flow direction;

[0133] Saline osmotic pressure π, dynamic viscosity μ and salt diffusion coefficient D s are calculated by the following fitting formula:

[0134] π = 4.54047 (10 3 C / M s ρ) 0.987 (28)

[0135] μ = (1.4757 × 10 -3 + 2.4817 × 10 -6 C + 9.3287 × 10 -9 C 2 ) exp (-0.02008T) (29)

[0136] D s = 6.725 × 10 -6 exp (0.1546 × 10 -3 C - 2513 / (T + 273.15)) (30)

[0137] where M s is the molar mass of solute;

[0138] Step 2. Establishing reverse osmosis superstructure model;

[0139] The reverse osmosis system comprises seawater intake and pretreatment, product water post-treatment, a reverse osmosis membrane group, a pump, a pressure exchanger (PX), a stream mixer and a separator. Since there are carbonate ions in seawater, precipitates are easily formed during reverse osmosis operation. The seawater is acidified to adjust the pH value to 4.0, and the free CO2 in the seawater is removed by a blower. The decarbonized seawater can effectively reduce the scaling tendency. The pH value is increased to alkalinity by adding a strong base to effectively increase the boron rejection rate. The reverse osmosis superstructure comprises N PS pressurization stages and N RO reverse osmosis stages, and a total of N PS +2 stream nodes, 2 refers to the brine and fresh water finally leaving the reverse osmosis system, and N PS of the stream nodes each represent a stream that is pressurized by a high-pressure pump (or not pressurized without a high-pressure pump) and directly enters a reverse osmosis unit. Each reverse osmosis stage is composed of a plurality of parallel pressure vessels, and each pressure vessel is composed of 2-8 membrane elements connected in series and operating under the same conditions. Each stream of brine and fresh water leaving the reverse osmosis stage can enter N PS +2 stream nodes, each stream is represented as a function of flow rate, salinity, boron concentration and pressure. Each feed M IN passes through a stream distributor and is divided into M OUT streams, and some streams are mixed into a stream by a stream mixer. The stream distributor and the mixer are represented as:

[0140]

[0141] C in，out =C in out=1,...,M OUT (32)

[0142] C TB，in，out =C TB，in out=1,...,M OUT (33)

[0143] P in，out =P in out=1,...,M OUT (34)

[0144] pH in，out =pH in out=1,...,M OUT (35)

[0145]

[0146] 0=(P in -P out )Q in，outin=1,...M IN (40)

[0147] Formulas (31)-(35) represent the logistics distributor, formulas (36)-(39) represent the logistics mixer, and formula (40) represents the isobaric mixing constraint, which allows the reverse osmosis stage water to be mixed with the final system water and the reverse osmosis stage high-pressure brine to be depressurized and mixed with the system feed. The total boron concentration C TB Boric acid molecule C boric and borate ion C borate The sum of concentrations, Q in 、C in 、C TB，in and P in represent the flow rate, salinity, boron concentration and pressure at the inlet of the logistics distributor, respectively, Q out 、C out 、C TB，out and P out represent the flow rate, salinity, boron concentration and pressure at the outlet of the logistics mixer, respectively. Q in，out represents the outlet flow of the logistics distributor, C in，out 、C TB，in，out and P in，out represents the salinity, boron concentration and pressure at the outlet of the logistics distributor, Q acid and Q base are the amounts of acid and base added, respectively, C acid and C base are the concentrations of acid and base, respectively, C H + ，acid and C OH - ，base are the hydrogen ion concentrations in the acid and alkali solutions, respectively, and the subscripts in and out represent the inlet and outlet, respectively;

[0148] The material balance equation of the high-pressure pump and the work exchanger is:

[0149] Q ps，1 =Q hpp +Q pxlin (41)

[0150] Q ps，1 C ps，1 =Q hpp C hpp +Q pxlin C pxlin (42)

[0151] Q ps，1 C TB，ps，1 =Q hpp C TB，hpp +Q pxlinC TB，pxlin (43)

[0152] Q RO，1 =Q hpp +Q pxhout (44)

[0153] Q RO，1 C RO，1 =Q hpp C hpp +Q pxhout C pxhout (45)

[0154] Q RO，1 C TB，RO，1 =Q hpp C TB，hpp +Q pxhout C TB，pxhout (46)

[0155] Q pxhout =Q pxlin (47)

[0156] Q pxhin =Q pxlout (48)

[0157] L px Q pxhin / 100=Q pxhin -Q pxhout (49)

[0158] L px [%]=0.3924+0.01238P pxhin (50)

[0159] C pxhout =Mix(C pxhin -C pxlin )+C pxlin (51)

[0160] C TB，pxhout =Mix(C TB，pxhin -C TB，pxlin )+C TB，pxlin (52)

[0161] Mix=6.0057-0.3559OF+0.0084OF 2 (53)

[0162] OF[%]=100×(Q pxhin ,-Q pxhout ) / Q pxhin (54)

[0163] Cpxlout Q pxlout =Q pxlin C pxlin +Q pxhin C pxhin -Q pxhout C pxhout (55)

[0164] C TB，pxlout Q pxlout =Q pxlin C TB，pxlin +Q pxhin C TB，pxhin -Q pxhout C TB，pxhout (56)

[0165] Among them L PX is the leakage rate, Mix is ​​the volume mixing ratio, OF (-10% ≤ OF ≤ 15%) is the lubrication flow rate, the subscripts hpp, pxhin, pxlin, pxhout, and pxhin represent the high-pressure pump, the low-pressure feed seawater and high-pressure brine entering the work exchanger, and the pressurized seawater and decompressed brine leaving the work exchanger, respectively; the subscript ps represents the boosting stage, and RO represents the reverse osmosis stage;

[0166] The material flow leaving the i-th boosting stage directly enters the j-th reverse osmosis stage. The same type of membrane elements are used in the reverse osmosis pressure vessel of the same stage. Its characteristics such as pure water permeability coefficient, solute permeability coefficient, membrane area and feed spacer thickness remain unchanged. The membrane element model k used in the j-th reverse osmosis pressure vessel is determined by the following formula:

[0167]

[0168] Introduce binary variable y j，k When it is 1, it means that the kth type of membrane element is selected in the jth stage of reverse osmosis, otherwise it is 0; Formula (58) defines the maximum allowable water inlet pressure of the membrane element, U is a sufficiently large number, K t It is a collection of types of reverse osmosis membrane elements;

[0169] Formulas (60) and (61) are introduced to determine the brine and product water of the reverse osmosis stage entering the next reverse osmosis stage. If the binary variable β b，i，j =1, the brine of the jth stage enters the next reverse osmosis stage. p，i，j If it is equal to 1, the produced water of the jth stage enters the next reverse osmosis stage. Formula (62) can prevent the concentrated brine and produced water from entering the next reverse osmosis stage at the same time.

[0170]

[0171] As water passes through the membrane, the pH of the brine and product water will also change, which is described by the following formula:

[0172]

[0173] in represents the recovery rate of the j-th stage reverse osmosis, r l，j It represents the recovery rate of each differential unit in the j-th stage reverse osmosis pressure vessel, It can be determined that the feed seawater of the jth reverse osmosis stage is the previous reverse osmosis stage or reverse osmosis section;

[0174] The pH value of seawater is 8.2. To effectively remove CO2 from seawater, use (Q acid，0 +Q acid，1 ) strong acid to adjust the pH value of seawater to 4.0, and then remove the CO2 in the seawater through the purge tower, and then add (Q base，0 +Q base，1 ) to increase the pH value of seawater to increase the boron removal rate.

[0175]

[0176] (Q f +Q acid，0 )10 -7 +Q acid，1 ·C acid =(Q f +Q acid，0 +Q acid，1 )10 -4 (69)

[0177] (Q f +Q acid，0 +Q acid，1 )10 -4 =Q base，0 C base (70)

[0178]

[0179] Wherein formula (71) calculates the amount of strong base required to increase the pH value of the reverse osmosis stage feed;

[0180] The entire reverse osmosis network satisfies the following material balance relationship and product water demand constraints:

[0181] Q f =Q b +Q p (72)

[0182] Q f C f =Qb C b +Q p C p (73)

[0183] Q f C TB，f =Q b C TB，b +Q p C TB，p (74)

[0184]

[0185] Q p ≥Q p，lo (81)

[0186] C p ≤C p.up (82)

[0187] C TB，p ≤C TB，p，up (83)

[0188] Where Q b 、C b and C TB，b are the brine flow rate, salinity and boron concentration leaving the reverse osmosis network, Q p 、C p and C TB，p represent the flow rate, salinity and boron concentration of product water, respectively, and the subscripts lo and up represent the minimum required value and the maximum allowed value, respectively;

[0189] Step 3. System flow and operating condition constraints;

[0190] To ensure the safe operation of the reverse osmosis system, the following constraints are set in the model: the concentration polarization factor is the salt concentration C on the membrane surface. ch.mw.1 and the salinity of the bulk solution C ch，b.1 The concentration polarization factor limit of the first-stage reverse osmosis of decarbonized seawater is 1.22, the concentration polarization factor limit of the first-stage reverse osmosis of traditional seawater feed is 1.2, and the concentration polarization factor of the second-stage reverse osmosis is 1.4 at most because the salt content of its feed water has been significantly reduced; the maximum pressure loss of a single pressure vessel is 0.35 MPa, and the maximum average water production flux of the first and second stages is 20 L / (m 2 h) and 40L / (m 2 ·h), the maximum water flux of the first membrane element of the first stage and the second stage is 35L / (m 2 h) and 48L / (m 2 ·h), the minimum brine flow rate in the first and second stage pressure vessels is 3.6m 3 / h and 2.4m3 / h, the brine concentration is less than 90kg / m 3 , the maximum pH value of reverse osmosis feed is 9.5, and the maximum pH value of reverse osmosis grade feed is 11.0:

[0191] Step 4. Establish a reverse osmosis system optimization design model

[0192] The optimization design problem of the reverse osmosis system is expressed as a mixed integer nonlinear programming, with formula (79) as the objective function and satisfying the process thermodynamics, unit operation, and design requirements constraints:

[0193]

[0194] The objective function Ew takes into account the system energy consumption, total water intake, membrane module scale and acid and alkali reagent addition, ΔP SWIP , ΔP hpp , ΔP bp and ΔP bppx They represent the pressure difference of seawater intake pump, high pressure pump, interstage pump and booster pump respectively, Q f , Q hpp , Q bp and Q bppx They represent the flow rates of feed seawater, high-pressure pump, interstage pump and booster pump respectively, η is the efficiency of the equipment, f c is the load factor of the reverse osmosis device, n m，j and n pv，j represents the number of membrane elements in the j-th stage reverse osmosis pressure vessel and the number of the j-th stage reverse osmosis pressure vessel, and the subscript moter is the motor of the pump;

[0195] Step 5. Solve the formed reverse osmosis system optimization problem;

[0196] The optimization objective is Ew formula (84),

[0197] The constraint equations are: reverse osmosis process model formulas (1)-(30),

[0198] Reverse osmosis superstructure model formulas (31)-(83),

[0199] System flow and operating condition constraints;

[0200] A two-level stochastic model is used to optimize the process and operating conditions of the reverse osmosis system under uncertain conditions. The uncertainty parameters in the stochastic optimization model are expressed as probability distribution Ω = {ω1, ω2}, so the two-level stochastic model is expressed as:

[0201]

[0202] constraint:

[0203]

[0204] constraint:

[0205] The first two equations represent the first-level model, and the last two equations represent the second-level model. and are decision variables, including the flow rate Q, pressure P, pH at each level in the reverse osmosis model, and the logistics distributor outlet variable Q in，out , represents the number of membrane elements n in the j-th level reverse osmosis pressure vessel m，j , the number of reverse osmosis pressure vessels in the jth stage n pvj , binary variable Y l and binary variable y j，k ;c T are the parameters in the objective function, is the mathematical expectation of the optimal solution of the second-level model. Formula (85) represents the constraint equation in the optimization problem of the reverse osmosis system, where are the coefficients of the constraint equations, and is the right side of the constraint equation, T ω and W ω are the transposed and compensation matrices respectively; in the first-level model, n m，j and n pv，j Holding constant, other decision variables are adjusted for uncertainty in the second-level model;

[0206] Mathematical programming software is used to solve the above mixed integer nonlinear programming problem. By assigning different initial values ​​to the variables and iterating from multiple initial points, the optimized process and operating conditions of the system can be obtained.

[0207] The present invention is described in detail below with reference to the embodiments:

[0208] The present invention conducts a case study on a reverse osmosis seawater desalination system with a spiral membrane element. The salinity and boron content of seawater are 35 kg / m 3 and 0.005kg / m 3 Table 1 lists the basic parameters of the SW membrane element for seawater desalination and the BW membrane element for brackish water desalination, and Table 2 lists the relevant parameters of the optimization model. To solve this optimization problem, the reverse osmosis pressure vessel was divided into 30 finite difference nodes. The DE solver of the General Algebraic Modeling System (GAMS) software was used to solve the stochastic optimization problem, and the SBB solver was used to solve the resulting mixed-integer nonlinear programming problem.

[0209] Table 1. Basic parameters of membrane elements

[0210] Membrane element types SW BW <![CDATA[有效膜面积[m 2 ]]]> 37.2 40.9 Membrane element length [m] 1.016 1.016 Membrane element diameter [m] 0.201 0.201 Effective diameter of membrane element [m] 0.88 0.88 Cross-sectional area of the water inlet channel S fcs [m 2 ]]]> 0.0150 0.0165 Feed screen height h[m] <![CDATA[7.112×10 -4 ]]> <![CDATA[7.112×10 -4 <!-- 15 -->]]> <![CDATA[隔网孔隙率ε sp ]]> 0.9 0.9 <![CDATA[进水流道当量直径,d e [m]]]> <![CDATA[8.126×10 -4 ]]> 8.126 x 10 -4 ]] <![CDATA[进料流量范围[m 3 / h]]]> 0.8-16 0.8-17 Maximum operating pressure [Mpa] 8.3 4.1 Continuous operating pH range 2-11 2-11 <![CDATA[纯水透过系数A ref [kg / m 2 ·s·Pa]]]> <![CDATA[3.5×10 -9 ]]> <![CDATA[1.128×10 -8 ]]> <![CDATA[盐透过系数B ref [m / s]]]> <![CDATA[3.2×10 -8 ]]> <![CDATA[4.421×10 -8 ]]> <![CDATA[B(OH)3透过系数B boric，ref [m / s]]]> 1.667 x 10 -6 ]] 7.813 x 10 -6 ]] <![CDATA[B(OH)4 - Transmission coefficient B borate，ref [m / s]]]> <![CDATA[9.803×10 -8 ]]> <![CDATA[9.387×10 -8 ]]>

[0211] Table 2. RO optimization model parameters

[0212] <![CDATA[平均浓盐水密度ρ[kg / m 3 ]]]> 1020 Universal gas constant R [J / (mol·K)] 8.314 Solute molecular weight Ms 58.5 <![CDATA[取水泵出口压力P swip [MPa]]]> 0.5 <![CDATA[取水 / 高压 / 级间 / 增压泵效率η swip / or bpp / or bp / or bpxp ]]> 75％ Heat exchanger efficiency η px ]]> 95％ Motor efficiency η motor ]]> 98％ <![CDATA[反渗透运行载荷系数f c ]]> 0.9 coefficient of friction, K λ ]]> 2.4 Strong base / strong acid concentration [mol / L] 1.474 / 0.984

[0213] Attachment Figure 1 Schematic diagram of a reverse osmosis deboronation superstructure for decarbonized seawater feed. Seawater is pumped from intake pump 1 to pretreatment 2. After strong acid 3 is added to adjust the pH to 4.0, carbon dioxide 6 is removed by blower 5 in decarbonization tower 4. After strong base 7 is added to the decarbonized seawater, it is pumped to primary mixer 8 and then split into two streams. One stream is pressurized by high-pressure pump 9 and enters mixer 11. The other stream enters work exchanger 10, where it is pressurized by booster pump 12 and also transported to mixer 11. The mixed seawater enters primary reverse osmosis 15. The reverse osmosis produced water and concentrated brine enter the logistics distributor and mixer respectively. The concentrated brine or reverse osmosis produced water passes through the secondary mixer 19 and is pressurized by the interstage pump 13 before entering the secondary reverse osmosis 14. The concentrated brine of the secondary reverse osmosis is depressurized by the pressure reducing valve 16 and returned to the primary mixer. The produced water flow rate at the front end of each level of reverse osmosis pressure vessel is regulated by the flow regulating valve 17 and 18 respectively. Strong alkali 20 can be added after the interstage pump 13 to adjust the pH value. Logistics 21 and 22 are the final produced water and final concentrated brine respectively.

[0214] Attachment Figure 1 This is a schematic diagram of a conventional reverse osmosis seawater desalination and boron removal superstructure. Seawater is transported to the pretreatment unit 24 by the water intake pump 23. After strong acid or strong base is added to the primary mixer 25, it is divided into two streams. One stream is pressurized by the high-pressure pump 26 and enters the mixer 28, and the other stream enters the work exchanger 27 and is pressurized by the booster pump 29 and is also transported to the mixer 28. The mixed seawater enters the primary reverse osmosis 31. The reverse osmosis product water and the concentrated brine enter the logistics distributor and mixer respectively. The concentrated brine or reverse osmosis product water passes through the secondary mixer 35 and is pressurized by the interstage pump 36 and then enters the secondary reverse osmosis 30. The concentrated brine of the secondary reverse osmosis is released by the pressure reducing valve 34 and returned to the primary mixer 25. The water production flow rate at the front end of each level of reverse osmosis pressure vessel is regulated by the flow regulating valve 32 and 33 respectively. Strong base 37 can be added after the interstage pump to adjust the pH value. Logistics 38 and 39 are the final product water and the final concentrated brine, respectively.

[0215] The common model parameters in the example are as follows:

[0216] Reverse osmosis water production: 120m 3 / h; Maximum allowable water salinity: 0.50kg / m 3 ; Annual membrane water flux attenuation: 7% for the first stage and 2% for the second stage; Annual membrane salt rejection rate and boron rejection rate increase: 10% for the first stage and 5% for the second stage.

[0217] The experimental data reported in the literature were used to verify the decarbonized seawater reverse osmosis model proposed in this paper. The experimental conditions were as follows: the first-stage reverse osmosis feed seawater pressure was 5 MPa, the pressure drop was 0.02 MPa, the pH value was 9.3, and the salinity was 36.953 kg / m 3 , temperature 25℃; the secondary reverse osmosis feed seawater is the primary reverse osmosis product water, pressure 1.1MPa, pressure drop 0.02MPa, pH 9.6, salinity 0.772kg / m 3 , temperature 25℃, as attached Figure 3 , Attachment Figure 4 , Attachment Figure 5 , Attachment Figure 6 and attached Figure 7 As shown in the figure, the predicted results of boron concentration and pH value of the produced water from the first-stage reverse osmosis and the second-stage reverse osmosis are in good agreement with the experimental values.

[0218] In Example 1, the contamination coefficient is set as an uncertainty parameter, and its probability distribution and optimization results are shown in Table 3. The number of membrane elements in the first, second and third stage reverse osmosis pressure vessels are 7, 8 and 5 respectively. The whole process is as follows Figure 8 As shown in a, the whole process covers all process designs in different scenarios. According to the different pollution coefficients, the reverse osmosis system operates in different scenarios. The process in different scenarios is adjusted according to the actual situation. For example, the process in scenario 1 is Figure 8 b. Scenario 2 and Scenario 3 use process 8c. During operation, the number of pressure vessels in operation is adjusted accordingly with different scenarios. For Scenario 3 with the lowest pollution coefficient, due to the more serious pollution situation, the feed pressure needs to be increased to 6.42MPa. The number of pressure vessels for brackish water membrane elements in operation is large, so the energy consumption increases significantly. Table 4 shows the conventional seawater optimization scheme. The whole process of the system is as follows: Figure 9 As shown in Figure a, while this solution has a higher recovery rate than the decarbonized seawater optimization solution, due to the lower pH of the primary reverse osmosis feedwater and the lower boron removal rate, the number of membrane elements in the secondary and tertiary reverse osmosis stages is higher than that of the decarbonized seawater optimization solution, resulting in higher system energy consumption. Compared with the conventional seawater optimization solution, decarbonized seawater can save energy between 15.4% and 19.0%.

[0219] Table 3. Uncertainty optimization scheme for decarbonized seawater pollution coefficient

[0220]

[0221] Table 4. Uncertainty optimization scheme for conventional seawater pollution coefficient

[0222]

[0223] In Example 2, as shown in Table 4, the contamination coefficient and the maximum boron content of the produced water are set as uncertainty parameters. The reverse osmosis system produces produced water with different boron contents under different feed seawater conditions to meet different requirements. The optimization results show that the number of membrane elements in the first, second, and third stage reverse osmosis pressure vessels are 7, 6, and 6, respectively. The whole process is as follows: Figure 10 As shown in a, in different scenarios, some logistics or equipment (pumps, reverse osmosis stages) are shut down to reduce system energy consumption. When the boron content of the produced water is high (0.001 kg / m in scenarios 1, 4 and 7), the boron content is 0.001 kg / m 3 ) The first-stage reverse osmosis system is an energy-saving solution. For other scenarios, the second-stage reverse osmosis system can achieve a lower boron removal rate. When the boron concentration of the produced water is 0.0005kg / m 3 In scenarios 2 and 5, the third-stage reverse osmosis water is returned to the second stage for reprocessing. For scenario 8, where the pollution is more serious, a water diversion design is adopted. When the boron concentration of the produced water is 0.0003 kg / m 3 When the pollution coefficient is 0.85, the third-stage produced water needs to be transported to the previous stage for treatment. Due to the increase in the pH value of the first-stage reverse osmosis feed and the reasonable adjustment of the process, the system recovery rate is maintained at a high level of 48.95-53.56%, so the maximum system energy consumption is 3.232kWh / m 3 , the lowest can be as low as 2.639kWh / m 3 The conventional seawater optimization scheme is shown in Table 6, and the process is as follows Figure 11 As shown in the figure, conventional seawater optimization schemes for different scenarios all use a three-stage reverse osmosis optimization scheme, and their energy consumption is relatively high. Decarbonized seawater can save energy consumption between 14.5% and 21.0%.

[0224] Table 5. Optimization scheme for uncertainty of decarbonized seawater pollution coefficient and maximum boron concentration

[0225]

[0226] Table 6. Optimization scheme for conventional seawater pollution coefficient and maximum boron concentration uncertainty

[0227]

[0228] Case analysis shows that the use of decarbonized seawater can effectively reduce the energy consumption of the reverse osmosis system, and the use of a stochastic optimization model can take into account the uncertainty of the system during operation. When the pollution coefficient is an uncertain factor, the decarbonized seawater optimization scheme only needs to adjust the system operating parameters according to different scenarios to meet the desalination and deboronization requirements. When the pollution coefficient and the maximum boron content are both uncertain parameters, the system energy consumption can be reduced by adjusting the system process and operating parameters. Compared with the conventional optimization scheme, the two examples solved the energy consumption between 14.5% and 21.0%, which is of great significance for energy conservation and emission reduction.

[0229] The examples of the present invention have been specifically described above, but the present invention is not limited to the examples. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

[0230] Matters not covered by the present invention are known technologies.

Claims

1. A method for optimizing a decarbonized seawater feed reverse osmosis deboronation seawater desalination system considering uncertainty, characterized in that The method comprises the following steps: Step 1: Establish a spiral membrane element reverse osmosis desalination process model; According to the reverse osmosis process mechanism, differential equations are used to describe the axial changes of salinity, boron concentration, pressure, and flow in the pressure vessel. The finite difference method differential equation is discretized to obtain the following equation: C boric,l =α 0,l C TB,l (8) K boric，l =2K l (13) Among them, A, B, B TB are the permeability coefficients of pure water, salt and boron respectively, P represents pressure, C represents salinity, π represents osmotic pressure, ρ p and V w represent the density and flow rate of fresh water, J w and J s are pure water flux and salt permeation flux, V w is the permeation flow rate, T is the temperature, K is the mass transfer coefficient, d e is the equivalent diameter of the feed channel, S l is the area of ​​a differential unit of the membrane element, S l =S m ·n m / L,S m is the area of ​​a single membrane element, n m is the number of pressure vessel membrane elements, L is the total number of differential unit nodes, Re is the Reynolds number, Re=ρVd e / μ, where μ is the dynamic viscosity, Sc is the Schmidt number, Sc=μ / ρD s , D s is the diffusion coefficient of salt, Q is the flow rate, V is the feed flow rate, V = Q / (3600S fcs ε sp ), S fcs is the cross-sectional area of ​​the feed channel, ε sp is the porosity of the feed channel screen, σ is the reflection coefficient, α0 and α1 are the fractions of boric acid and borate ions, respectively, and pK a is the primary ionization constant of boric acid, K λ is the friction coefficient, FF d is the contamination coefficient, e is the activation energy of the membrane (when T≤298K, e=25,000J / mol -1 When T>298K, e=22,000J / mol -1 ), R is the gas constant, B in is the annual salt permeation increase rate, N mlp is the average life of the reverse osmosis membrane, N RO Represents the total reverse osmosis level, k λ represents the internal friction coefficient of the membrane element, the subscript ch is the feed or product water flow channel of the membrane element, b is concentrated brine, boric is boric acid molecule, borate is borate ion, f is feed seawater, p is product water, mw is membrane surface, ref is the membrane parameter when T0 is free of contamination at 298K, l is the differential unit node, and j represents the jth reverse osmosis stage; Boundary conditions for the finite difference method: z=0,V=V in ,Q=Q in ,C boric =C boric,in ,C borate =C borate,in ,C=C in ,P=P in ; Where in represents the pressure vessel inlet; The water production diversion design directly delivers the water produced at the front end of the pressure vessel to the final water production, and the water produced at the back end enters the next stage of reverse osmosis for desalination. Flow regulating valves are added at both ends of the pressure vessel to adjust the water output ratio at both ends. The model is expressed as follows: Q ch,b,l+1 =Q ch,b,l -3600V w，l S l (15) Q ch,b，l+1 C ch,b,l+1 -Q ch，b，l C ch，b，l =-3600V w,l S l C ch,p，l (16) Q ch,b,l+1 C TB,ch,b,l+1 -Q ch,b,l C TB,ch,b,l =-3600V w,l S l C TB,ch,p,l (17) Q f,n =Q b，n +Q p,n,lc +Q p,n,hc (24) Q f，n C f,n =Q b，n C b,n +Q p,n,lc C p,n,lc +Q p,n,hc C p,n,hc (25) Q f,n C TB,f,n =Q b，n C TB,b,n +Q p,n,lc C TB,p,n,lc +Q p,n,hc C TB，p，n，hc (26) AND l -AND l+1 ≥0 (27) Wherein the binary variable Y represents the water flow direction of the differential unit in the pressure vessel, the subscript lc represents the water produced at the front end of the pressure vessel, hc represents the water produced at the rear end of the pressure vessel, and n represents the nth pressure vessel. Formula (27) indicates that the water produced at the front end or rear end of each differential unit in the pressure vessel has a consistent flow direction. Salt water osmotic pressure π, dynamic viscosity μ and salt diffusion coefficient D s Calculated by the following fitting formula: π=4.54047(10 3 C / M s p) 0.987 (28) μ=(1.4757×10 -3 +2.4817×10 -6 C+9.3287×10 -9 C 2 )exp(-0.02008T) (29) D s =6.725×10 -6 exp(0.1546×10 -3 C-2513 / (T+273.15)) (30) Among them, M s is the molar mass of the solute; Step 2. Establish a reverse osmosis superstructure model; The reverse osmosis system includes seawater intake and pretreatment, water post-treatment, reverse osmosis membrane group, pump, pressure exchanger (PX), logistics mixer and separator. Due to the presence of carbonate ions in seawater, precipitation is easily formed during the reverse osmosis operation. The seawater is acidified and the pH value is adjusted to 4.

0. The free CO2 in the seawater is removed by the fan. Decarbonized seawater can effectively reduce the tendency of scaling. Adding strong alkali to increase the pH value to alkaline effectively increases the boron retention rate. The reverse osmosis superstructure contains N PS boost stages and N RO There are N reverse osmosis stages in total. PS +2 logistics nodes, 2 refers to the salt water and fresh water that eventually leave the reverse osmosis system, N PS Each of the logistics nodes represents a stream that enters a reverse osmosis unit directly after being pressurized by a high-pressure pump (or without being pressurized by a high-pressure pump). Each reverse osmosis stage is composed of multiple parallel pressure vessels, each pressure vessel consists of 2 to 8 membrane elements in series and operates under the same conditions. Each stream of salt water and fresh water leaving the reverse osmosis stage can enter N PS +2 logistics nodes, each logistics is expressed as a function of flow rate, salinity, boron concentration and pressure, each feed M IN After passing through the logistics distributor, it is divided into M OUT Some logistics are combined into one logistics after passing through the logistics mixer. The logistics distributor and mixer are expressed as: C in,out =C in out=1,...M OUT (32) C TB,in,out =C TB,in out=1,...M OUT (33) P in,out =P in out=1,...M OUT (34) pH in,out =pH in out=1,...M OUT (35) 0=(P in -P out )Q in,out in=1,...M IN (40) Formulas (31)-(35) represent the logistics distributor, formulas (36)-(39) represent the logistics mixer, and formula (40) represents the isobaric mixing constraint, which allows the reverse osmosis stage water to be mixed with the final system water and the reverse osmosis stage high-pressure brine to be depressurized and mixed with the system feed. The total boron concentration C TB Boric acid molecule C boric and borate ion C borate The sum of concentrations, Q in 、C in 、C TB,in and P in represent the flow rate, salinity, boron concentration and pressure at the inlet of the logistics distributor, respectively, Q out 、C out， C TB,out and P out represent the flow rate, salinity, boron concentration and pressure at the outlet of the logistics mixer, respectively. Q in,out represents the outlet flow of the logistics distributor, C in,out 、C TB,in,out and P in,out represents the salinity, boron concentration and pressure at the outlet of the logistics distributor, Q acid and Q base are the amounts of acid and base added, respectively, C acid and C base are the concentrations of acid and base, respectively, are the hydrogen ion concentrations in the acid and alkali solutions, respectively, and the subscripts in and out represent the inlet and outlet, respectively; The material balance equation of the high-pressure pump and the work exchanger is: Q ps，1 =Q hpp +Q pxlin (41) Q ps,1 C ps，1 =Q hpp C hpp +Q pxlin C pxlin (42) Q ps，1 C TB，ps，1 =Q hpp C TB,hpp +Q pxlin C TB,pxlin (43) Q RO，1 =Q hpp +Q pxhout (44) Q RO，1 C RO，1 =Q hpp C hpp +Q pxhout C pxhout (45) Q RO，1 C TB，RO，1 =Q hpp C TB,hpp +Q pxhout C TB,pxhout (46) Q pxhout =Q pxlin (47) Q pxhin =Q pxlout (48) L px Q pxhin / 100=Q pxhin -Q pxhout (49) L px [%]=0.3924+0.01238P pxhin (50) C pxhout =Mix(C pxhin -C pxlin )+C pxlin (51) C TB,pxhout =Mix(C TB,pxhin -C TB,pxlin )+C TB，pxlin (52) Mix=6.0057-0.3559OF+0.0084OF 2 (53) OF[%]=100×(Q pxhin ,-Q pxhout ) / Q pxhin (54) C pxlout Q pxlout =Q pxlin C pxlin +Q pxhin C pxhin -Q pxhout C pxhout (55) C TB,pxlout Q pxlout =Q pxlin C TB，pxlin +Q pxhin C TB,pxhin -Q pxhout C TB,pxhout (56) Among them L PX is the leakage rate, Mix is ​​the volume mixing ratio, OF (-10% ≤ OF ≤ 15%) is the lubrication flow rate, the subscripts hpp, pxhin, pxlin, pxhout, and pxhin represent the high-pressure pump, the low-pressure feed seawater and high-pressure brine entering the work exchanger, and the pressurized seawater and decompressed brine leaving the work exchanger, respectively; the subscript ps represents the boosting stage, and RO represents the reverse osmosis stage; The material flow leaving the i-th boosting stage directly enters the j-th reverse osmosis stage. The same type of membrane elements are used in the reverse osmosis pressure vessel of the same stage. Its characteristics such as pure water permeability coefficient, solute permeability coefficient, membrane area and feed spacer thickness remain unchanged. The membrane element model k used in the j-th reverse osmosis pressure vessel is determined by the following formula: P j -P k，max ≤U(1-y j,k ) j=1,2,...,N RO ,k=1,2,...,K t (58) Introduce binary variable y j，k When it is 1, it means that the kth type of membrane element is selected in the jth stage of reverse osmosis, otherwise it is 0; Formula (58) defines the maximum water inlet pressure allowed by the membrane element, U is a sufficiently large number, K t It is a collection of types of reverse osmosis membrane elements; Formulas (60) and (61) are introduced to determine the brine and product water of the reverse osmosis stage entering the next reverse osmosis stage. If the binary variable β b，i，j =1, the brine of the jth stage enters the next reverse osmosis stage. p，i,j If it is equal to 1, the produced water of the jth stage enters the next reverse osmosis stage. Formula (62) can prevent the concentrated brine and produced water from entering the next reverse osmosis stage at the same time. As water passes through the membrane, the pH of the brine and product water will also change, which is described by the following formula: in represents the recovery rate of the j-th stage reverse osmosis, r l,j It represents the recovery rate of each differential unit in the j-th stage reverse osmosis pressure vessel, It can be determined that the feed seawater of the jth reverse osmosis stage is the previous reverse osmosis stage or reverse osmosis section; The pH value of seawater is 8.

2. To effectively remove CO2 from seawater, use (Q acid,0 +Q acid，1 ) strong acid to adjust the pH value of seawater to 4.0, and then remove the CO2 in the seawater through the purge tower, and then add (Q base,0 +Q base,1 ) to increase the pH value of seawater to increase the boron removal rate. (Q f +Q acid,0 )10 -7 +Q acid,1 ·C acid =(Q f +Q acid,0 +Q acid,1 ]10 -4 (69) (Q f +Q acid,0 +Q acid,1 )10 -4 =Q base，0 C base (70) Wherein formula (71) calculates the amount of strong base required to increase the pH value of the reverse osmosis stage feed; The entire reverse osmosis network satisfies the following material balance relationship and product water demand constraints: Q f =Q b +Q p (72) Q f C f =Q b C b +Q p C p (73) Q f C TB，f =Q b C TB，b +Q p C TB,p (74) Q p ≥Q p，lo (81) C p ≤C p,up (82) C TB，p ≤C TB,p,up (83) Where Q b 、C b and C TB，b are the brine flow rate, salinity and boron concentration leaving the reverse osmosis network, Q p 、C p and C TB,p represent the flow rate, salinity and boron concentration of product water, respectively, and the subscripts lo and up represent the minimum required value and the maximum allowed value, respectively; Step 3. System flow and operating condition constraints; To ensure the safe operation of the reverse osmosis system, the following constraints are set in the model: the concentration polarization factor is the salt concentration C on the membrane surface. ch.mw.l and the salinity of the bulk solution C ch,b.l The concentration polarization factor limit of the first-stage reverse osmosis of decarbonized seawater is 1.22, and the concentration polarization factor limit of the first-stage reverse osmosis of traditional seawater feed is 1.

2. The concentration polarization factor of the second-stage reverse osmosis is 1.4 at most because the salt content of its feed water has been significantly reduced. The maximum pressure loss of a single pressure vessel is 0.35 MPa, and the maximum average water production flux of the first and second stages is 20 L / (m 2 h) and 40L / (m 2 ·h), the maximum water flux of the first membrane element of the first stage and the second stage is 35L / (m 2 h) and 48L / (m 2 ·h), the minimum brine flow rate in the first and second stage pressure vessels is 3.6m 3 / h and 2.4m 3 / h, the brine concentration is less than 90kg / m 3 , the maximum pH value of reverse osmosis feed is 9.5, and the maximum pH value of reverse osmosis grade feed is 11.0; Step 4. Establish a reverse osmosis system optimization design model The optimization design problem of the reverse osmosis system is expressed as a mixed integer nonlinear programming, with formula (79) as the objective function and satisfying the process thermodynamics, unit operation, and design requirements constraints: The objective function Ew takes into account the system energy consumption, total water intake, membrane module scale and acid and alkali reagent addition, ΔP SWIP , ΔP hpp , ΔP bp and ΔP bppx They represent the pressure difference of seawater intake pump, high pressure pump, interstage pump and booster pump respectively, Q f , Q hpp , Q bp and Q bppx They represent the flow rates of feed seawater, high-pressure pump, interstage pump and booster pump respectively, η is the efficiency of the equipment, f c is the load factor of the reverse osmosis device, n m，j and n pv，j represents the number of membrane elements in the j-th stage reverse osmosis pressure vessel and the number of the j-th stage reverse osmosis pressure vessel, and the subscript moter is the motor of the pump; Step 5. Solve the formed reverse osmosis system optimization problem; The optimization objective is Ew formula (84), The constraint equations are: reverse osmosis process model formulas (1)-(30), Reverse osmosis superstructure model formulas (31)-(83), System flow and operating condition constraints; A two-level stochastic model is used to optimize the process and operating conditions of the reverse osmosis system under uncertain conditions. The uncertainty parameters in the stochastic optimization model are expressed as probability distribution Ω = {ω1, ω2}, so the two-level stochastic model is expressed as: constraint: constraint: The first two equations represent the first-level model, and the last two equations represent the second-level model. and are decision variables, including the flow rate Q, pressure P, pH at each level in the reverse osmosis model, and the logistics distributor outlet variable Q in,out , represents the number of membrane elements in the j-th level reverse osmosis pressure vessel n m,j , the number of reverse osmosis pressure vessels in the jth stage n pv,j , binary variable Y l and binary variable y j,k ;c T are the parameters in the objective function, is the mathematical expectation of the optimal solution of the second-level model. Formula (85) represents the constraint equation in the optimization problem of the reverse osmosis system, where are the coefficients of the constraint equations, and is the right side of the constraint equation, T ω and W ω are the transposed and compensation matrices respectively; in the first-level model, n m，j and n pv,j Holding constant, other decision variables are adjusted for uncertainty in the second-level model; Mathematical programming software is used to solve the above mixed integer nonlinear programming problem. By assigning different initial values ​​to the variables and iterating from multiple starting points, the optimized process and operating conditions of the system can be obtained.

Images