Energy-integrated heat-coupled multi-component distillation process structure design method for flash recompression of bottom liquid

Abstract

The present invention discloses a method for designing the structure of a multi-component distillation process with energy-integrated thermal coupling of a bottom liquid-phase flash recompression. First, the distillation process is structurally represented to obtain a distillation process structure; then the design problem of a distillation process that simultaneously includes a distillation sequence structure, a bottom liquid-phase flash recompression structure, a thermal coupling structure, and an energy-integrated structure is converted into an optimization problem; then, the SA-PSO double-layer optimization framework is used to sequentially perform PSO optimization and SA optimization on any N-component mixture to be separated to obtain an optimal solution. The present invention implements a complete representation of a multi-component mixture to be separated that simultaneously includes a distillation sequence, a bottom liquid-phase flash recompression, a thermal coupling, and an energy-integrated structure, then establishes an implicit MINLP model with TAC as the objective function, converts the design problem of the distillation process into an optimization problem, and finally uses the SA-PSO stochastic optimization framework that combines the SA algorithm with the PSO algorithm to achieve simultaneous optimization of discrete variables and continuous variables, and completes an efficient solution.

Description

Technical Field

[0001] The present invention belongs to the field of chemical process system engineering (PSE for short), and relates to an optimization focusing on a distillation system, and specifically is an energy-integrated heat-coupled multi-component distillation process structure design method for flash evaporation and recompression of liquid phase in a bottom. Background Art

[0002] Distillation is the main separation method in the chemical and petroleum industries due to its advantages of large processing capacity and simple operation. However, the traditional distillation process often consumes a high proportion of energy in the chemical process and has a thermodynamic efficiency of less than 10%.

[0003] Distillation sequence synthesis is an important means to achieve energy saving in the distillation process. In order to improve the energy utilization efficiency of the distillation process, a variety of energy-saving technologies, such as non-clear separation, flash recompression of the bottom liquid phase, thermal coupling and energy integration technology, need to be considered simultaneously in the sequence synthesis.

[0004] For multi-component (component number ≥ 3) mixtures to be separated, the large-scale combinatorial characteristics and strong non-convex characteristics presented by the distillation sequence synthesis problem will bring great difficulty to the solution of the model. Up to now, the existing technology has not been able to achieve the design and optimization of distillation sequences for multi-component mixtures to be separated while considering non-clear separation, flash recompression of the bottom liquid phase, thermal coupling and energy integration technology. Summary of the invention

[0005] In view of the deficiencies in the prior art, the technical problem to be solved by the present invention is to provide an energy-integrated heat-coupled multi-component distillation process structure design method for flash evaporation and recompression of liquid phase in the bottom of a tower.

[0006] The technical solution of the present invention to solve the technical problem is to provide an energy-integrated heat-coupled multi-component distillation process structure design method for liquid phase flash recompression in the bottom of the tower, characterized in that the method comprises the following steps:

[0007] Step 1: Confirm the initial information and objective function;

[0008] The initial information includes the feed conditions of the mixture to be separated, the physical property parameters of each component in the mixture, the cold utility conditions used, the hot utility conditions used, the electricity price and the value range of the set decision variables;

[0009] The objective function is the TAC of the entire distillation process;

[0010] Step 2, structurally represent the distillation process to obtain a distillation process structure; the distillation process structure consists of a distillation sequence structure, a bottom liquid flash recompression structure, a thermal coupling structure and an energy integration structure;

[0011] Step 3, converting the design problem of the distillation process including the distillation sequence structure, the bottom liquid flash recompression structure, the thermal coupling structure and the energy integration structure into an optimization problem;

[0012] The TAC calculation of a randomly generated distillation process is shown in formula (3):

[0013] C TAC =β·C cap ({Tree},{TC},p,N t ,R,F,ξ LK ,ξ HK ,CR,Q C ,Q R ,T D ,T B )+C op ({Tree},{TC},Q util ) (3)

[0014] In formula (3), the integer array {Tree} represents the distillation sequence, the 0-1 array {TC} represents the thermal coupling structure; p, N t , R, F, ξ LK , HK , CR are the operating pressure, theoretical plate number, actual reflux ratio, feed rate, recovery rate of light key components, recovery rate of heavy key components and compression ratio of the bottom liquid flash recompression structure of all distillation towers in the distillation process; Q C and T D are the heat load and temperature of all condensers in the distillation process; Q R and T B are the heat load and temperature of all reboilers in the distillation process; Q util It represents the utilities consumed by the distillation process, including heat utilities, cold utilities and electricity;

[0015] In order to ensure that the TAC of a feasible distillation process that meets the separation requirements is given, all variables in formula (3) must satisfy the three constraints of material balance, energy balance and thermodynamic equation;

[0016] Based on the FUG shortcut method, R and N t can be expressed as ξ LK , HK and r, where r is the actual reflux ratio R and the minimum reflux ratio R min Based on material balance, F depends on {Tree}, ξ LK and HK ; Based on energy balance, Q C and Q R Depends on {Tree}, p, r, ξLK and HK ; Based on bubble point calculation, T D Depends on p, {Tree}, ξ LK , HK and CR, T B Depends on p, {Tree}, ξ LK and HK ; Energy integration strategy based on pinch point method and assumption of constant minimum heat transfer temperature difference, Q util Depends on p, {Tree}, ξ LK , HK and CR, and then transform equation (3) into equation (4):

[0017] C TAC =C cap ({Tree},{TC},p,r,ξ LK ,ξ HK ,CR)+C op ({Tree},{TC},p,r,ξ LK ,ξ HK ,CR)

[0018] =C cost ({Tree},{TC},p,r,ξ LK ,ξ HK ,CR) (4)

[0019] Step 4: Build a SA-PSO two-layer optimization framework for any N-component mixture to be separated;

[0020] Step 5: Use the SA-PSO two-layer optimization framework built in step 4 to perform PSO optimization and SA optimization on any N-component mixture to be separated in turn to obtain the optimal solution of this optimization;

[0021] Step 5.1, for the N-component mixture to be separated, randomly generate a binary tree to represent the distillation sequence structure {Tree}, and set the initial temperature T0 of the annealing temperature T of the outer iterative SA algorithm;

[0022] Step 5.2: For the generated distillation sequence structure {Tree}, it is necessary to determine whether it can be converted into an N-1 tower distillation process; if not, the distillation sequence structure {Tree} is updated by using the method of generating adjacent solutions using the SA algorithm until the generated distillation sequence structure {Tree} can be converted into an N-1 tower distillation process, and then a set of 0-1 variables is randomly generated to represent the thermal coupling structure {TC} of the N-1 tower distillation process;

[0023] Step 5.3, the distillation sequence structure {Tree} that satisfies the N-1 tower distillation process enters the inner layer for PSO optimization: based on the distillation sequence structure {Tree} generated in step 5.2 and the corresponding thermal coupling structure {TC}, the positions and velocities of all particles are initialized; for each particle, the design calculation of each distillation tower is performed based on the FUG shortcut method, and the tower bottom liquid phase flash recompression structure is determined by the value of CR; after obtaining the temperature and heat flow rate of the cold and hot streams, the energy integration structure is obtained by pinch point analysis, and the value of the particle fitness is calculated, which is the value of the objective function TAC;

[0024] Step 5.4: Update the positions and velocities of all particles in PSO until the number of iterations t reaches the maximum number of iterations t max , the iteration of the inner optimization stops, the distillation sequence structure {Tree} and the thermal coupling structure {TC} and the optimized p, r, ξ LK ,ξ HK ,CR five continuous variables are recorded as new solutions;

[0025] Step 5.5: The new solution enters the outer layer for SA optimization: First, the TAC of the current solution and the optimal solution in the first iteration is set to be large enough, and then the Metropolis criterion is used to determine whether to accept the new solution; if accepted, the new solution replaces the current solution and the process proceeds to step 5.6; if not, the new solution is discarded and the current optimal solution is retained;

[0026] Step 5.6: Determine whether to update {Tree} and its corresponding {TC} in the current optimal solution at the same time or only update {TC} to ensure that the better {Tree} can be optimized at different initial positions and speeds in the PSO inner optimization: make judgments with equal probability in the early stage of the SA process; in the later stage of the SA process, update {Tree} and generate corresponding {TC} with a probability of 70-90%, and do not update {Tree} and only update {TC} with a probability of 10-30%;

[0027] Step 6: After a complete PSO optimization and SA optimization, the annealing temperature T is reduced by the cooling formula of the SA algorithm; when the annealing temperature T does not drop to the termination temperature T end , if step 5.6 updates the {Tree} of the current solution and its corresponding {TC}, then repeat steps 5.2 to 5.6; when the annealing temperature T does not drop to the termination temperature T end If step 5.6 only updates {TC}, then repeat steps 5.3 to 5.6; when the annealing temperature T drops to the termination temperature T end When , the iteration ends, the optimal solution is obtained and all the information of the optimal solution is output.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] (1) The present invention realizes a complete representation of the distillation sequence, bottom liquid flash recompression, thermal coupling and energy integration structure for multi-component mixtures to be separated. Then, an implicit MINLP model with TAC as the objective function is established to transform the design problem of the distillation process into an optimization problem. Finally, the SA-PSO stochastic optimization framework combining the SA algorithm and the PSO algorithm is used to realize the simultaneous optimization of discrete variables and continuous variables, thereby achieving an efficient solution.

[0030] (2) The present invention generates a distillation sequence structure through a binary tree coding method, which can achieve a complete expression of clear separation and unclear separation. The compression ratio CR is then introduced to represent the flash recompression structure of the bottom liquid phase, and a 0 / 1 variable is used to represent the thermal coupling structure. Finally, the energy integration structure is designed based on the principle of pinch point analysis.

[0031] (3) The present invention adopts an electrically driven bottom liquid flash recompression structure and integrates the stream after flash with the top steam, thereby saving the use of the bottom heat utility, thus promoting the energy transformation of the distillation process. In addition, the introduction of non-clear separation and thermal coupling technology will further reduce the energy consumption of the distillation process.

[0032] (4) The present invention has the characteristics of fast solution speed, good convergence quality, good stability and high computational efficiency. It can obtain an optimal solution with good quality and a series of suboptimal solutions with a high probability. The optimal solution is of great significance for energy conservation and emission reduction, and the heat energy consumption is 0, which means less fossil energy use and lower CO2 emissions. In addition, the TAC of each suboptimal solution is similar to that of the optimal solution, and can also be used as an alternative for industrial applications.

[0033] (5) The present invention realizes efficient comprehensive design and optimization of large-scale complex distillation sequences through the implementation of structural representation, mathematical modeling and stochastic optimization strategies, and the results can be applied to the preliminary design stage of multi-component mixture distillation systems in actual industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a flow chart of the core part of the present invention;

[0035] Figure 2 It is a schematic diagram of coding of each component of the mixture to be separated of the present invention;

[0036] Figure 3 A schematic diagram of a separation task and an equivalent binary tree node structure of the present invention;

[0037] Figure 4 It is a schematic diagram of the thermally coupled stream feed of the present invention;

[0038] Figure 5 This is a schematic diagram of product stream feed of the present invention;

[0039] Figure 6 It is a schematic diagram of energy matching based on pinch point analysis of the present invention;

[0040] Figure 7 This is an iterative curve diagram of five-component alcohols in Example 1 of the present invention;

[0041] Figure 8 This is the optimal solution obtained in Example 1 of the present invention;

[0042] Fig. 9 This is the second optimal solution obtained in Example 1 of the present invention;

[0043] Fig.10 This is the third optimal solution obtained in Example 1 of the present invention;

[0044] Fig.11 This is the fourth optimal solution obtained in Example 1 of the present invention;

[0045] Fig.12 It is the fifth optimal solution obtained in Example 1 of the present invention;

[0046] Fig.13 This is the sixth optimal solution obtained in Example 1 of the present invention. DETAILED DESCRIPTION

[0047] The specific embodiments of the present invention are given below. The specific embodiments are only used to further illustrate the present invention in detail and do not limit the protection scope of the claims of the present invention.

[0048] The present invention provides an energy-integrated heat-coupled multi-component distillation process structure design method for liquid-phase flash recompression in a tower bottom (hereinafter referred to as the method), characterized in that the method comprises the following steps:

[0049] Step 1: Confirm the initial information and objective function;

[0050] The initial information includes the feed conditions of the mixture to be separated, the physical property parameters of each component in the mixture, the cold utility conditions used, the hot utility conditions used, the electricity price and the value range of the set decision variables;

[0051] The objective function is the TAC (total annual cost) of the entire distillation process, and its function value is used to describe the quality of the design; the smaller the TAC, the better the design;

[0052] Preferably, in step 1, the TAC calculation formula is expressed as:

[0053] C TAC =β·C cap +C op (1)

[0054] In formula (1), C capis the equipment cost function, which is the equipment cost function of all distillation columns, condensers, reboilers and heat exchangers required for the energy matching of hot and cold streams in the distillation process, with the unit of $; C op is the annual operating cost function, including the annual hot utility cost, cold utility cost and electricity cost function in the process, with the unit of $ / year; β is the annual depreciation factor, with the unit of year -1 .

[0055] Step 2: Represent the structure of the distillation process to obtain the distillation process structure; the distillation process structure is composed of a distillation sequence structure, a flash re-compression structure of the bottom liquid phase, a heat integration structure and an energy integration structure;

[0056] Preferably, step 2 is specifically: in the process of randomly generating the distillation process structure, first generate the distillation sequence structure by binary tree coding, then introduce the compression ratio CR to represent the flash re-compression structure of the bottom liquid phase, and use 0 / 1 variables to represent the heat integration structure; design the energy integration structure based on the principle of pinch analysis to obtain the distillation process structure.

[0057] Preferably, in step 2, the binary tree coding method is used to generate the distillation sequence structure, specifically: sort the components in the N-component mixture to be separated in descending order of volatility and number them from 1 to N (as Figure 2 shown, N≥3), and then any mixed stream can be represented by a two-dimensional array composed of the codes of the heaviest and lightest components; [m,n] represents a mixed stream of n-m+1 components, where m and n are the codes of the lightest and heaviest components in the mixed stream respectively; and the separation task of separating the mixed stream of [m,n] can be represented by a binary tree (as Figure 3 shown); by randomly assigning integers x and y under the constraint conditions m≤x<n, m<y≤n and y≤x+1, any feasible separation task k is generated;

[0058] Preferably, in step 2, the calculation formula of the compression ratio CR is as follows:

[0059]

[0060] In formula (2), p c is the compressor outlet pressure, and p i is the compressor inlet pressure;

[0061] For the separation task k, when the obtained value of CR k is greater than 1, it indicates that the flash re-compression structure of the bottom liquid phase exists; when the value of CR k is equal to 1, it indicates that the flash re-compression structure of the bottom liquid phase does not exist.

[0062] Preferably, in step 2, the 0 / 1 variable represents a thermal coupling structure, specifically: there are two ways to connect two separation tasks in any distillation sequence: thermally coupled streams (including liquid and gas streams with opposite flow directions, such as Figure 4 ) or from a product stream produced from a condenser or reboiler (such as Figure 5 where 0 represents saturated liquid feed and 1 represents thermally coupled stream feed.

[0063] Preferably, in step 2, the energy integration structure is designed based on the principle of pinch point analysis, specifically: above the pinch point, the energy of each hot stream is matched in order of the hot stream temperature from high to low; below the pinch point, the energy of each cold stream is matched in order of the cold stream temperature from low to high. Figure 6 As shown, above the pinch, hot stream a is first matched with cold stream f for heat exchange, and then hot stream b is matched with cold stream g for heat exchange; below the pinch, cold stream q is first matched with hot stream d for heat exchange, and then cold stream p is matched with hot stream d for heat exchange; thereby reducing the use of high-grade cold and hot utilities.

[0064] Step 3, converting the design problem of the distillation process including the distillation sequence structure, the bottom liquid flash recompression structure, the thermal coupling structure and the energy integration structure into an optimization problem;

[0065] The TAC calculation of a randomly generated distillation process is shown in formula (3):

[0066] C TAC =β·C cap ({Tree},{TC},p,N t ,R,F,ξ LK ,ξ HK ,CR,Q C ,Q R ,T D ,T B )+C op ({Tree},{TC},Q util ) (3)

[0067] In formula (3), the integer array {Tree} represents the distillation sequence, the 0-1 array {TC} represents the thermal coupling structure; p, N t , R, F, ξ LK , HK , CR are the operating pressure, theoretical plate number, actual reflux ratio, feed rate, recovery rate of light key components, recovery rate of heavy key components and compression ratio of the bottom liquid flash recompression structure of all distillation towers in the distillation process; Q C and T Dare the heat load and temperature of all condensers in the distillation process; Q R and T B are the heat load and temperature of all reboilers in the distillation process; Q util It represents the utilities consumed by the distillation process, including heat utilities, cold utilities and electricity;

[0068] In order to ensure that the TAC of a feasible distillation process that meets the separation requirements is given, all variables in formula (3) must satisfy the three constraints of material balance, energy balance and thermodynamic equation.

[0069] Based on the FUG (Fenske-Underwood-Gilliland) shortcut method, R and N t can be expressed as ξ LK , HK and r, where r is the actual reflux ratio R and the minimum reflux ratio R min Based on material balance, F depends on {Tree}, ξ LK and HK ; Based on energy balance, Q C and Q R Depends on {Tree}, p, r, ξ LK and HK ; Based on bubble point calculation, T D Depends on p, {Tree}, ξ LK , HK and CR, T B Depends on p, {Tree}, ξ LK and HK ; Energy integration strategy based on pinch point method and assumption of constant minimum heat transfer temperature difference, Q util Depends on p, {Tree}, ξ LK , HK and CR, and then transform equation (3) into equation (4):

[0070] C TAC =C cap ({Tree},{TC},p,r,ξ LK ,ξ HK ,CR)+C op ({Tree},{TC},p,r,ξ LK ,ξ HK ,CR)

[0071] =C cost ({Tree},{TC},p,r,ξ LK ,ξ HK ,CR) (4)

[0072] Preferably, in step 3, for formula (4), generally, the smaller the TAC is, the better the distillation process structure design is. Therefore, the design problem of the distillation process including the distillation sequence structure, the bottom liquid flash recompression structure, the thermal coupling structure and the energy integration structure is transformed into an optimization problem, and the corresponding implicit MINLP (mixed integer nonlinear programming) model is established, as shown in formula (5):

[0073]

[0074] In formula (5), C cost is the objective function TAC; {Tree},{TC},p,r,ξ LK ,ξ HK , CR is the decision variable of the distillation process structure; CSS represents all feasible distillation sequences, scp is the sequence encoding parameters and related constraints when updating {Tree}; TCS represents all thermal coupling structures, tcp is the thermal coupling structure parameters and related constraints when updating {TC}; the compressor outlet temperature of separation task k Should be greater than the dew point temperature of the compressed stream To prevent liquefaction from damaging the compressor; the compressor outlet temperature of separation task k It should also be less than the critical temperature of the compressed stream Avoid the formation of supercritical fluid; K represents a series of separation tasks in the distillation sequence.

[0075] Step 4: Build a SA-PSO (simulated annealing algorithm-particle swarm optimization algorithm) two-layer optimization framework for any N-component mixture to be separated on the Visual Studio platform and compile it using C++ language;

[0076] Preferably, in step 4, in the SA-PSO two-layer optimization framework, the SA algorithm optimizes the discrete variables {Tree} and {TC}; the PSO algorithm optimizes the continuous variables p, r, ξ LK ,ξ HK and CR.

[0077] Step 5: Use the SA-PSO two-layer optimization framework built in step 4 to perform PSO optimization and SA optimization on any N-component mixture to be separated in turn to obtain the optimal solution of this optimization;

[0078] Step 5 is as follows:

[0079] Step 5.1, for the N-component mixture to be separated, randomly generate a binary tree to represent the distillation sequence structure {Tree}, and set the initial temperature T0 of the annealing temperature T of the outer iterative SA algorithm;

[0080] Step 5.2: For the generated distillation sequence structure {Tree}, it is necessary to determine whether it can be converted into an N-1 tower distillation process; if not, the distillation sequence structure {Tree} is updated by using the method of generating adjacent solutions using the SA algorithm until the generated distillation sequence structure {Tree} can be converted into an N-1 tower distillation process, and then a set of 0-1 variables is randomly generated to represent the thermal coupling structure {TC} of the N-1 tower distillation process;

[0081] Step 5.3, the distillation sequence structure {Tree} that satisfies the N-1 tower distillation process enters the inner layer for PSO optimization: based on the distillation sequence structure {Tree} generated in step 5.2 and the corresponding thermal coupling structure {TC}, the positions and velocities of all particles are initialized; for each particle, the design calculation of each distillation tower is performed based on the FUG shortcut method, and the tower bottom liquid phase flash recompression structure is determined by the value of CR; after obtaining the temperature and heat flow rate of the cold and hot streams, the energy integration structure is obtained by pinch point analysis, and the value of the particle fitness is calculated, which is the value of the objective function TAC;

[0082] Step 5.4: Update the positions and velocities of all particles in PSO until the number of iterations t reaches the maximum number of iterations t max , the iteration of the inner optimization stops, the distillation sequence structure {Tree} and the thermal coupling structure {TC} and the optimized continuous variables (i.e. p, r, ξ LK ,ξ HK ,CR) is recorded as a new solution;

[0083] Preferably, in step 5.4, the particle position and velocity update formulas are as follows:

[0084]

[0085]

[0086] In formula (6), particles represent a corresponding set of continuous variables: operating pressure p, light key component recovery rate ξ LK 、Recovery rate of heavy key componentsξ HK , actual reflux ratio R and minimum reflux ratio R min The ratio r and compression ratio CR of the particle; the particle position corresponds to the value of the continuous variable, and the particle speed corresponds to the direction of change of the continuous variable value; in the D-dimensional solution space, each particle i has a position x i (x i1 ,x i2 ,…,x ij ,…,x iD ) and a velocity v i (v i1 ,v i2 ,…,v ij,…,v iD ); the optimal position of particle i is denoted as pbest i , the optimal position of the entire population is recorded as gbest; t represents the current iteration number of the PSO algorithm; c1 and c2 represent learning factors and their values ​​are both 2; r1 and r2 represent random numbers between 0 and 1; ω represents the inertia factor, which reflects the change of the particle itself to the current velocity. The calculation formula is as follows:

[0087]

[0088] In formula (7), t max It represents the maximum number of iterations of PSO. The value of ω increases with the number of PSO iterations from the initial inertia factor ω ini To the final inertia factor ω end Linear decrease;

[0089] Step 5.5: The new solution enters the outer layer for SA optimization: First, the TAC of the current solution and the optimal solution in the first iteration is set to be large enough, and then the Metropolis criterion is used to determine whether to accept the new solution; if accepted, the new solution replaces the current solution and the process proceeds to step 5.6; if not, the new solution is discarded and the current optimal solution is retained;

[0090] Step 5.6, determine whether to update {Tree} and its corresponding {TC} in the current optimal solution at the same time or only update {TC} to ensure that the better {Tree} can be optimized at different initial positions and speeds in the PSO inner optimization: in the early stage of the SA process, make a judgment with equal probability; in the later stage of the SA process, since the generated inferior solution will be rejected with a greater probability (the greater probability is 1-PP), update {Tree} and generate the corresponding {TC} with a probability of 70-90% (preferably 80%), and do not update {Tree} and only update {TC} with a probability of 10-30% (preferably 20%);

[0091] Preferably, in step 5.5, the Metropolis criterion represents the probability of accepting a new solution to replace the current solution, which is divided into two cases: TAC new ≤TAC old , the new solution is accepted and replaced by the new solution; while TAC new >TAC old , then the new solution is an inferior solution, and the probability PP of accepting the inferior solution needs to be calculated. The calculation formula is as follows:

[0092]

[0093] In formula (8), TAC new Indicates the TAC value of the new solution; TAC oldrepresents the TAC value of the old solution; e is the natural base, which is 2.718; T represents the annealing temperature. As the annealing temperature T decreases, the probability of the inferior solution being accepted decreases.

[0094] Step 6: After a complete PSO optimization and SA optimization, the annealing temperature T is reduced by the cooling formula of the SA algorithm; when the annealing temperature T does not drop to the termination temperature T end , if step 5.6 updates the {Tree} of the current solution and its corresponding {TC}, then repeat steps 5.2 to 5.6; when the annealing temperature T does not drop to the termination temperature T end If step 5.6 only updates {TC}, then repeat steps 5.3 to 5.6; when the annealing temperature T drops to the termination temperature T end When , the iteration ends, the optimal solution is obtained and all the information of the optimal solution is output.

[0095] Preferably, in step 6, the cooling formula of the SA algorithm is as follows:

[0096] T j+1 =αT j (9)

[0097] In formula (9), T j represents the annealing temperature of the jth iteration, T j+1 represents the annealing temperature of the j+1th iteration, α (0<α<1) represents the temperature reduction coefficient, which is 0.95-0.995.

[0098] Example

[0099] In this implementation, the separation of a five-component alcohol system is taken as an example. The purpose of this embodiment is to optimize the decision variables contained in the distillation process under the conditions of known five-component alcohol system feed information, public engineering, etc., so as to obtain the distillation process structure and operating parameters that meet the separation requirements and minimize the objective function value (TAC). At the same time, in order to ensure the quality of the solution obtained, it is usually necessary to repeat the program 300-500 times to screen the optimal solution and the suboptimal solution.

[0100] Step 1: In this embodiment, the number of components is 5, the raw material stream is a saturated liquid mixture of ethanol, isopropanol, n-propanol, isobutanol and n-butanol, the feed pressure is 0.1 MPa, and the total flow rate is 500.4 kmol / h. Specific initial information is shown in Tables 1-3:

[0101] Table 1 Feed conditions

[0102] Components Mole fraction Relative volatility (A) Ethanol 0.25 4.6156 (B) Isopropyl alcohol 0.15 4.0046 (C) n-Propanol 0.35 2.2454 (D) Isobutanol 0.10 1.4775 (E) n-butanol 0.15 1.0000

[0103] Table 2 Public engineering conditions

[0104] Public works Temperature (K) Unit price ($ / GJ) Condensation 303 0.354 Heating steam 1 433 13.28 Heating steam 2 457 14.19 electricity — 16.8

[0105] Table 3 Decision variable settings

[0106]

[0107]

[0108] In step 1, the useful life of the equipment is set to 10 years, and the annual depreciation coefficient of the equipment is β 0.1year -1 The annual operating cost of the distillation process can be calculated based on the amount of cold and hot utilities and electricity consumed in the process operation for one year and the corresponding prices. In this embodiment, the operating time of the distillation process is 8000h / year.

[0109] In step 2, the initial information is input in the form of an array or a constant, so that the automatic representation of the distillation sequence structure, the bottom liquid flash recompression structure, the thermal coupling structure and the energy integration structure can be realized.

[0110] In step 5.1, the initial temperature T0 = 30K;

[0111] In step 5.3, set the total number of particles to 100;

[0112] In step 5.4, the maximum number of iterations t max Set to 20; ω ini =0.9,ω end =0.4;

[0113] In step 5.7, in the early stage of the SA process (T≥0.1K), the judgment is made with equal probability; in the late stage of the SA process (T<0.1K), {Tree} is updated with a probability of 80% and the corresponding {TC} is generated, and {Tree} is not updated and only {TC} is updated with a probability of 20%;

[0114] In step 6, the temperature reduction coefficient α is set to 0.995, and the end temperature T end Set to 0.001K.

[0115] The test results of this embodiment show that the present invention has a faster solution speed and good convergence quality. To solve this problem once, the outer layer simulated annealing is iterated 2056 times, and the average CPU computing time required is 448.8 seconds. Figure 7 The horizontal axis in the middle represents the number of outer SA iterations, and the vertical axis represents the TAC of the current optimal solution. As the number of iterations increases, the TAC of the current optimal solution will decrease rapidly and eventually approach a flat level.

[0116] Figure 8-Figure 13The optimal solution, second optimal solution, third optimal solution, fourth optimal solution, fifth optimal solution and sixth optimal solution obtained in this embodiment are shown respectively, and the probability of obtaining the above solutions is 63.4%, 24.8%, 2.4%, 7.0%, 2.0% and 0.4% respectively. Compared with the optimal solution, the TAC of each suboptimal solution increases by 0.02%, 2.29%, 3.33%, 3.47% and 3.99% respectively. This embodiment can obtain the optimal solution and each suboptimal solution with better quality with a higher probability, and the TAC of each suboptimal solution is similar to that of the optimal solution, which can also be used as an alternative solution for industrial applications.

[0117] by Figure 8 For example, Figure 8 There are 6 distillation towers in the process, and towers 2, 5 and 6 are merged to form an N-1 tower process. There are thermally coupled streams ABC, AB, CD and DE at the top of tower 1, the top of the merged tower, the top of tower 3 and the bottom of tower 3 respectively, and there is a tower bottom liquid flash recompression structure at the bottom of tower 1, the top of tower 4 and the bottom of the merged tower. In addition, the two circles connected by dotted lines represent the energy integration structure formed.

[0118] The results obtained in the present invention can be verified by a strict model in Aspen Plus, and show good performance in application tests in the chemical industry such as performance analysis and result analysis. The present invention also has good effects on the sequence synthesis of six-component, seven-component and other systems with more components.

[0119] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A method for designing a structure of an energy-integrated heat-coupled multi-component distillation process for flash recompression of liquid phase in the bottom of a tower, characterized in that: The method comprises the following steps: Step 1: Confirm the initial information and objective function; The initial information includes the feed conditions of the mixture to be separated, the physical property parameters of each component in the mixture, the cold utility conditions used, the hot utility conditions used, the electricity price and the value range of the set decision variables; The objective function is the TAC of the entire distillation process; Step 2, structurally represent the distillation process to obtain a distillation process structure; the distillation process structure consists of a distillation sequence structure, a bottom liquid flash recompression structure, a thermal coupling structure and an energy integration structure; Step 3, converting the design problem of the distillation process including the distillation sequence structure, the bottom liquid flash recompression structure, the thermal coupling structure and the energy integration structure into an optimization problem; The TAC calculation of a randomly generated distillation process is shown in formula (3): C TAC =β·C cap ({Tree},{TC},p,N t ,R,F,ξ LK ,ξ HK ,CR,Q C ,Q R ,T D ,T B )+C op ({Tree},{TC},Q util ) (3) In formula (3), the integer array {Tree} represents the distillation sequence, the 0-1 array {TC} represents the thermal coupling structure; p, N t , R, F, ξ LK , HK , CR are the operating pressure, theoretical plate number, actual reflux ratio, feed rate, recovery rate of light key components, recovery rate of heavy key components and compression ratio of the bottom liquid flash recompression structure of all distillation towers in the distillation process; Q C and T D are the heat load and temperature of all condensers in the distillation process; Q R and T B are the heat load and temperature of all reboilers in the distillation process; Q util Represents the utilities consumed by the distillation process, including heat utilities, cold utilities and electricity; C cap is the equipment cost function; C op is the annual operating cost function; β is the annual depreciation coefficient; In order to ensure that the TAC of a feasible distillation process that meets the separation requirements is given, all variables in formula (3) must satisfy the three constraints of material balance, energy balance and thermodynamic equation; Based on the FUG shortcut method, R and N t can be expressed as ξ LK , HK and r, where r is the actual reflux ratio R and the minimum reflux ratio R min Based on material balance, F depends on {Tree}, ξ LK and HK ; Based on energy balance, Q C and Q R Depends on {Tree}, p, r, ξ LK and HK ; Based on bubble point calculation, T D Depends on p, {Tree}, ξ LK , HK and CR, T B Depends on p, {Tree}, ξ LK and HK ; Energy integration strategy based on pinch point method and assumption of constant minimum heat transfer temperature difference, Q util Depends on p, {Tree}, ξ LK , HK and CR, and then transform equation (3) into equation (4): Step 4: Build a SA-PSO two-layer optimization framework for any N-component mixture to be separated; Step 5: Use the SA-PSO two-layer optimization framework built in step 4 to perform PSO optimization and SA optimization on any N-component mixture to be separated in turn to obtain the optimal solution of this optimization; Step 5.1, for the N-component mixture to be separated, randomly generate a binary tree to represent the distillation sequence structure {Tree}, and set the initial temperature T0 of the annealing temperature T of the outer iterative SA algorithm; Step 5.2: For the generated distillation sequence structure {Tree}, it is necessary to determine whether it can be converted into an N-1 tower distillation process; if not, the distillation sequence structure {Tree} is updated by using the method of generating adjacent solutions using the SA algorithm until the generated distillation sequence structure {Tree} can be converted into an N-1 tower distillation process, and then a set of 0-1 variables is randomly generated to represent the thermal coupling structure {TC} of the N-1 tower distillation process; Step 5.3, the distillation sequence structure {Tree} that satisfies the N-1 tower distillation process enters the inner layer for PSO optimization: based on the distillation sequence structure {Tree} generated in step 5.2 and the corresponding thermal coupling structure {TC}, the positions and velocities of all particles are initialized; for each particle, the design calculation of each distillation tower is performed based on the FUG shortcut method, and the tower bottom liquid phase flash recompression structure is determined by the value of CR; after obtaining the temperature and heat flow rate of the cold and hot streams, the energy integration structure is obtained by pinch point analysis, and the value of the particle fitness is calculated, which is the value of the objective function TAC; Step 5.4: Update the positions and velocities of all particles in PSO until the number of iterations t reaches the maximum number of iterations t max , the iteration of the inner optimization stops, the distillation sequence structure {Tree} and the thermal coupling structure {TC} and the optimized p, r, ξ LK ,ξ HK ,CR five continuous variables are recorded as new solutions; Step 5.5: The new solution enters the outer layer for SA optimization: First, the TAC of the current solution and the optimal solution in the first iteration is set to be large enough, and then the Metropolis criterion is used to determine whether to accept the new solution; if accepted, the new solution replaces the current solution and the process proceeds to step 5.6; if not, the new solution is discarded and the current optimal solution is retained; Step 5.6: Determine whether to update both the {Tree} and its corresponding {TC} in the current optimal solution or only update the {TC} to ensure that a better {Tree} can be optimized at different initial positions and velocities in the inner optimization of PSO. Make the determination with equal probability in the early stage of the SA process. In the later stage of the SA process, update the {Tree} and generate the corresponding {TC} with a probability of 70 - 90%, and do not update the {Tree} and only update the {TC} with a probability of 10 - 30%. Step 6: After a complete PSO optimization and SA optimization, the annealing temperature T is reduced by the cooling formula of the SA algorithm; when the annealing temperature T does not drop to the termination temperature T end , if step 5.6 updates the {Tree} of the current solution and its corresponding {TC}, then repeat steps 5.2 to 5.6; when the annealing temperature T does not drop to the termination temperature T end If step 5.6 only updates {TC}, then repeat steps 5.3 to 5.6; when the annealing temperature T drops to the termination temperature T end When , the iteration ends, the optimal solution is obtained and all the information of the optimal solution is output.

2. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 1, characterized in that: Specifically, Step 2 is as follows: In the process of randomly generating the rectification process structure, first generate the rectification sequence structure through binary tree encoding, then introduce the compression ratio CR to represent the structure of the bottom liquid flash and recompression, and use 0 / 1 variables to represent the heat integration structure. Finally, design the energy integration structure based on the principle of pinch analysis to obtain the rectification process structure.

3. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 2, characterized in that: In Step 2, the binary tree encoding method is used to generate the rectification sequence structure. Specifically, the components in the N-component mixture to be separated are sorted in descending order of volatility and numbered from 1 to N. Then, any mixed stream can be represented as a two-dimensional array composed of the encodings of the heaviest and lightest components. [m, n] represents a mixed stream of n - m + 1 components, where m and n are the encodings of the lightest and heaviest components in the mixed stream, respectively. The separation task of separating the mixed stream [m, n] can be represented by a binary tree. By randomly assigning integers x and y under the constraint conditions m ≤ x < n, m < y ≤ n, and y ≤ x + 1, any feasible separation task k is generated.

4. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 3, characterized in that: In Step 2, the calculation formula for the compression ratio CR is as follows: In formula (2), p c is the compressor outlet pressure, p i is the compressor inlet pressure; For separation task k, when the CR obtained by optimization is k When the value of is greater than 1, it indicates that the bottom liquid flash recompression structure exists; when CR k When the value of is equal to 1, it indicates that the liquid phase flash recompression structure in the bottom of the tower does not exist.

5. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 2, characterized in that: In Step 2, 0 / 1 variables are used to represent the heat integration structure. Specifically, there are two connection methods for two separation tasks in any rectification sequence: a heat integration stream that simultaneously includes liquid and gas streams with opposite flow directions or a product stream produced from a condenser or a reboiler. Among them, 0 represents saturated liquid feed, and 1 represents heat integration stream feed.

6. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 2, characterized in that: In Step 2, the energy integration structure is designed based on the principle of pinch analysis. Specifically, above the pinch point, the energy of each hot stream is matched in descending order of the hot stream temperature. Below the pinch point, the energy of each cold stream is matched in ascending order of the cold stream temperature.

7. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 1, characterized in that: In Step 3, for Equation (4), the smaller the TAC, the better the design of the rectification process structure. Therefore, the design problem of the rectification process that simultaneously includes the rectification sequence structure, the bottom liquid flash and recompression structure, the heat integration structure, and the energy integration structure is transformed into an optimization problem, and the corresponding implicit MINLP model is established as shown in Equation (5). In formula (5), C cost is the objective function TAC; {Tree},{TC},p,r,ξ LK ,ξ HK , CR is the decision variable of the distillation process structure; CSS represents all feasible distillation sequences, scp is the sequence encoding parameters and related constraints when updating {Tree}; TCS represents all thermal coupling structures, tcp is the thermal coupling structure parameters and related constraints when updating {TC}; the compressor outlet temperature of separation task k Should be greater than the dew point temperature of the compressed stream To prevent liquefaction from damaging the compressor; Compressor outlet temperature for separation task k It should also be less than the critical temperature of the compressed stream Avoid the formation of supercritical fluid; K represents a series of separation tasks in the distillation sequence.

8. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 1, characterized in that: In step 4, in the SA-PSO two-level optimization framework, the SA algorithm optimizes the discrete variables {Tree} and {TC}; the PSO algorithm optimizes the continuous variables p, r, ξ LK ,ξ HK and CR.

9. The energy-integrated heat-coupled multi-component distillation process structure design method for bottom liquid flash recompression according to claim 1, characterized in that: In Step 5.4, the particle position and velocity update formulas are as follows: In formula (6), particles represent a corresponding set of continuous variables: operating pressure p, light key component recovery rate ξ LK 、Recovery rate of heavy key componentsξ HK , actual reflux ratio R and minimum reflux ratio R min The ratio r and compression ratio CR of the particle; the particle position corresponds to the value of the continuous variable, and the particle speed corresponds to the direction of change of the continuous variable value; in the D-dimensional solution space, each particle i has a position x i (x i1 ,x i2 ,…,x ij ,…,x iD ) and a velocity v i (v i1 ,v i2 ,…,v ij ,…,v iD ); the optimal position of particle i is denoted as pbest i , the optimal position of the entire population is recorded as gbest; t represents the current iteration number of the PSO algorithm; c1 and c2 represent learning factors and their values ​​are both 2; r1 and r2 represent random numbers between 0 and 1; ω represents the inertia factor, which reflects the change of the particle itself to the current velocity. The calculation formula is as follows: In formula (7), t max It represents the maximum number of iterations of PSO. The value of ω increases with the number of PSO iterations from the initial inertia factor ω ini To the final inertia factor ω end Linear decrease; In step 5.5, the Metropolis criterion represents the probability of accepting a new solution to replace the current solution, which is divided into two cases: TAC new ≤TAC old , the new solution is accepted and replaced by the new solution; while TAC new >TAC old , then the new solution is an inferior solution, and the probability PP of accepting the inferior solution needs to be calculated. The calculation formula is as follows: In formula (8), TAC new Indicates the TAC value of the new solution; TAC old represents the TAC value of the old solution; e is the natural base; T represents the annealing temperature; In Step 6, the cooling formula of the SA algorithm is as follows: T j+1 =αT j (9) In formula (9), T j represents the annealing temperature of the jth iteration, T j+1 represents the annealing temperature of the j+1th iteration, and α represents the cooling coefficient.

Images