Propylene rectifying tower multi-objective energy-saving optimization method based on state simulation and improved wolf pack algorithm

Abstract

The invention relates to a multi-target energy-saving optimization method for a propylene rectifying tower based on state simulation and an improved wolf pack algorithm, which comprises the following steps of: S1, on the basis of a rectification process balance-level steady-state model, carrying out production working condition state simulation on the propylene rectifying tower based on a PRO / II process simulation platform, and establishing a propylene rectifying tower state simulation model; s2, on the basis of the state simulation model of the propylene rectifying tower, carrying out sensitivity analysis of single parameters and combined parameters on the operation parameters of the propylene rectifying tower, and establishing a data relation table between the propylene yield of the propylene rectifying tower and the raw material feeding temperature as well as between the energy consumption of the propylene tower and the tower top reflux quantity; s3, performing model parameter identification on the propylene rectifying tower data relation table by adopting a multiple regression analysis method and a least square method, and establishing a propylene tower multi-target energy-saving optimization model; s4, setting penalty factors to construct penalty function items, and converting the constrained multi-objective optimization function of the propylene tower into an unconstrained multi-objective optimization function; s5, in order to improve the local search capability of the optimization algorithm, the convergence speed is accelerated, the algorithm is prevented from falling into local optimum, and a dynamic search factor is added to improve the optimal solution search capability of the wolf pack algorithm; s6, executing an improved multi-objective wolf pack algorithm to solve an unconstrained multi-objective energy-saving optimization function of the propylene rectifying tower; and S7, judging whether the current maximum number of iterations reaches an upper limit value or not, if so, outputting population non-dominated individuals as a propylene tower multi-target energy-saving optimization result, and otherwise, returning to S4 to continue iterative calculation.

Description

Technical Field

[0001] The invention relates to the field of energy-saving optimization of distillation tower devices, and in particular to a multi-objective energy-saving optimization method for a propylene distillation tower based on state simulation and an improved wolf pack algorithm. Background Art

[0002] Petrochemical industry is a pillar industry of the national economy. According to statistics, the energy consumption of petrochemical industry accounts for about 15% of the total industrial energy consumption in the country. In petrochemical plants, the energy consumption of the distillation process accounts for about 60% of the total energy consumption. Therefore, to reduce the energy consumption of petrochemical plants, it is necessary to save energy in the distillation process, which is of great significance to the reduction of energy consumption of the entire petrochemical plant.

[0003] The distillation process is a complex mass transfer and heat transfer process involving multiple process variables, controlled variables and manipulated variables. Its process dynamics and mechanisms are also very complex. As one of the most widely used separation processes in chemical production, distillation is also a chemical unit operation with high energy consumption. However, in actual production, in order to ensure product quality, the operation of the distillation unit is often conservative, the operation method and parameter setting are usually not reasonable, and the phenomenon of over-separation is common. The propylene tower is the core device of the gas fractionation unit, which is mainly used to separate liquefied petroleum gas into high-purity propylene and propane fractions. High-purity propylene is the main product of the propylene tower. Due to the small difference in boiling point and volatility between propylene and propane, and the high separation purity requirement for propylene, the current propylene distillation tower device generally has high energy consumption. Therefore, the development of coordination and optimization between multiple goals based on the current production conditions of the distillation unit with the purpose of energy saving and efficiency improvement has always been the focus of the energy-saving optimization research of the distillation unit. Summary of the invention

[0004] The object of the present invention is to provide a multi-objective coordination and optimization method capable of optimizing the output of a propylene distillation tower and reducing the energy consumption of the device.

[0005] To achieve the above invention objectives, the present invention provides a multi-objective energy-saving optimization method for a propylene distillation tower based on state simulation and an improved wolf pack algorithm, comprising the following steps: S1. Based on the equilibrium-level steady-state model of the distillation process, the Peng-Robinson (PR) equation combined with the Inside-out convergence algorithm was used to simulate the production state of the propylene distillation tower based on the PRO / II process simulation platform, and a state simulation model of the propylene distillation tower was established; S2. Based on the state simulation model of the propylene distillation tower, the sensitivity analysis of single parameters and combined parameters of the operating parameters of the propylene distillation tower: tower top pressure, raw material feed temperature, tower top reflux rate is performed to establish a data relationship table between the propylene production of the propylene distillation tower and the energy consumption of the propylene tower and the raw material feed temperature and tower top reflux rate; S3. Use multiple regression analysis and least squares method to identify model parameters of the propylene distillation tower data relationship table, establish single-objective optimization functions of propylene tower propylene production and propylene tower energy consumption, and establish a multi-objective energy-saving optimization model for the propylene tower in combination with the actual operating parameter constraints of the propylene distillation tower; S4. According to formula (1), let the penalty factor be r k , construct the penalty function term Establish propylene production penalty function and propylene tower energy consumption penalty function, and transform the constrained propylene tower multi-objective optimization into unconstrained multi-objective optimization; In the formula, u is the number of initial iterations, n To terminate the iterations, is the objective function to be optimized, is the inequality constraint, express 0 takes the larger value. When x satisfies the constraint, =0, penalty term =0, when x does not satisfy the constraint condition, >0, penalty term >0, penalty factor r k+1 = cr k , k = k +1 gradually increases with the iteration of the algorithm, increasing the penalty for the optimization objective function; S5. In order to improve the local search ability of the optimization algorithm, speed up the convergence speed, and avoid the algorithm falling into the local optimum, a dynamic search factor is added θ Improve the ability of the wolf pack algorithm to search for the optimal solution, and calculate according to formula (2) θ The current value of In the formula, , is the dynamic search factor, t , t_max The algorithm is the current number of iterations and the maximum number of iterations; S6. Execute the improved multi-objective wolf pack algorithm to solve the unconstrained multi-objective energy-saving optimization function of the propylene distillation tower, including the following sub-steps: S6.1: After calculating the multi-objective optimization function value of the propylene tower (maximizing propylene production and minimizing propylene tower energy consumption) of each individual solution in the wolf pack algorithm, the current optimal solution is non-dominated and sorted according to the population size. N W Upper limit update Archive Optimal solution population; S6.2: Calculation parameters αand algorithm swing factor A , C The value of the parameter is shown in formula (3): In the formula, α The control parameter takes values ​​in the range [0,2]. r 1. r 2 is a random number in the range [0,1], t , t_max is the current number of iterations and the maximum number of iterations; S6.3: Roulette wheel selection method is used in Archive Select the optimal solution of the two optimization objective functions of the current propylene tower propylene production and the total energy consumption of the propylene distillation tower from the optimal solution population α 、Optimal solution β , suboptimal solution δ , and update the optimal position of individual wolves in the population according to formulas (2), (4), and (5), as shown in formulas (4) and (5): In the formula, The updated search direction of the wolf pack algorithm; S6.4: Calculate the objective function value of the new iterative search direction of the wolf pack individuals, that is, the current propylene production and energy consumption of the propylene tower. Construct a penalty function according to formula (1) to impose penalties on solutions that deviate from the operating parameter constraints of the propylene distillation tower. After updating the current propylene production and energy consumption optimization objective function values, perform non-dominated sorting on the wolf pack individual solutions to select the non-dominated solution of the current population. Archive The population individuals are compared and the dominated individual solutions are deleted to complete the pairing. Archive Population update.

[0006] S6.5: Judgment Archive Whether the population has reached the upper limit N W If the limit is exceeded, calculate the current Archive The crowding value distribution of the optimization objective function value (propylene tower output, propylene tower total energy consumption) of each individual in the group is calculated, and the individual solutions with too dense crowding are eliminated.

[0007] S7. Determine whether the current maximum number of iterations has reached the upper limit t_max , if yes, then output the non-dominated individuals in the population as the multi-objective optimization result of the propylene tower, otherwise return to S4 to continue the iterative calculation.

[0008] The present invention improves the original multi-objective wolf pack optimization algorithm, adds a penalty function mechanism and a dynamic search factor into the algorithm to solve the optimization model, avoids falling into a local optimal solution during the algorithm iterative search process, and can solve a combination of multiple groups of operating parameters, so that the propylene distillation tower can increase the production of propylene components while reducing the total energy consumption of the device, thereby realizing multi-objective energy-saving optimization of the propylene distillation tower device. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 . Schematic diagram of the method framework flow chart; Figures 2 to 8 .Sensitivity analysis of single operating parameters of propylene distillation tower; Figures 9 and 10 .Sensitivity analysis of combined parameters of propylene distillation tower; Fig.11 The data relationship between propylene production, raw material feed temperature and tower top reflux rate; Fig.12 Data relationship between propylene tower energy consumption, raw material feed temperature, and tower top reflux rate; Fig.13 Multi-objective energy-saving optimization results of propylene distillation tower; Fig.14 Multi-objective energy-saving optimization scheme and optimization results. DETAILED DESCRIPTION In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In the embodiments of the present invention, a multi-objective energy-saving optimization method for a propylene distillation tower based on state simulation and improved wolf pack algorithm is proposed, comprising the steps of: S1. Based on the equilibrium-level steady-state model of the distillation process, the production state of the propylene distillation tower was simulated by using the Peng-Robinson (PR) equation combined with the Inside-out convergence algorithm based on the PRO / II process simulation platform, and a state simulation model of the propylene distillation tower was established. Table 1 shows the state simulation model results and prediction accuracy of the propylene distillation tower. Table 1 Comparison between state simulation data and actual production data of propylene distillation unit Process parameters Design Specifications Simulation value Measured data Tower top pressure / Mpa 1.78 1.71 1.72 Tower top temperature / ℃ 44.2 41.1 44.1 Tower bottom temperature / ℃ 46.8 53.7 53.7 Propylene content at the top of the tower / % ≥99.5 99.5 99.5 Propane content at the bottom of the tower / % ≥95 95 95.01 Reflux ratio 16.4 18.79 19.22 S2. Figures 2 to 10 As shown in the figure, based on the state simulation model of the propylene distillation tower, the sensitivity analysis of single parameters and combined parameters of the operating parameters of the propylene distillation tower: tower top pressure, raw material feed temperature, tower top reflux volume is carried out, and the data relationship table of propylene production and propylene tower energy consumption of the propylene distillation tower with raw material feed temperature and tower top reflux volume is established. The specific data relationship is shown in Fig.11 , Fig.12 As shown; S3. Use multiple regression analysis and least squares method to identify model parameters of the propylene distillation tower data relationship table, and establish a single-objective optimization function for propylene tower propylene production and propylene tower energy consumption. The specific form is:

[0011] In the formula: x1 is the reflux rate of the propylene tower (t / h), x2 is the feed temperature (℃); z1 is the propylene production (t / h), z2 is the condenser load (M*kJ / h), and z3 is the reboiler load (M*kJ / h).

[0012] Combined with the actual operating parameter constraints of the propylene distillation tower, a multi-objective optimization model for the propylene tower is established. The specific form is:

[0013]

[0014] Where: x1 is the reflux rate of the propylene tower (t / h), x2 is the feed temperature (°C); y1 is the propylene output (t / h), y2 is the energy consumption of the propylene tower (M*kJ / h); S4. According to formula (1), let the penalty factor be r k , construct the penalty function term Establish propylene production penalty function and propylene tower energy consumption penalty function, and transform the constrained propylene tower multi-objective optimization into unconstrained multi-objective optimization; In the formula, u is the number of initial iterations, n To terminate the iterations, is the objective function to be optimized, is the inequality constraint, express 0 takes the larger value. When x satisfies the constraint, =0, penalty term =0, when x does not satisfy the constraint condition, >0, penalty term >0, penalty factor r k+1 = cr k , k = k +1 gradually increases with the iteration of the algorithm, increasing the penalty for the optimization objective function; S5. In order to improve the local search ability of the optimization algorithm, speed up the convergence speed, and avoid the algorithm falling into the local optimum, a dynamic search factor is added θ Improve the ability of the wolf pack algorithm to search for the optimal solution, and calculate according to formula (2) θ The current value of In the formula, , is the dynamic search factor, t , t_max The algorithm is the current number of iterations and the maximum number of iterations; S6. Execute the improved multi-objective wolf pack algorithm to solve the unconstrained multi-objective energy-saving optimization function of the propylene distillation tower, including the following sub-steps: S6.1: After calculating the multi-objective optimization function value of the propylene tower (maximizing propylene production and minimizing propylene tower energy consumption) of each individual solution in the wolf pack algorithm, the current optimal solution is non-dominated and sorted according to the population size. N W Upper limit update Archive Optimal solution population; S6.2: Calculation parameters α and algorithm swing factor A , C The value of the parameter is shown in formula (3): In the formula, α The control parameter takes values ​​in the range [0,2]. r 1. r 2 is a random number in the range [0,1], t , t_max is the current number of iterations and the maximum number of iterations; S6.3: Roulette wheel selection method is used in Archive Select the optimal solution of the two optimization objective functions of the current propylene tower propylene production and the total energy consumption of the propylene distillation tower from the optimal solution population α 、Optimal solution β , suboptimal solution δ , and update the optimal position of individual wolves in the population according to formulas (2), (4), and (5), as shown in formulas (4) and (5): In the formula, The updated search direction of the wolf pack algorithm; S6.4: Calculate the objective function value of the new iterative search direction of the wolf pack individuals, that is, the current propylene production and energy consumption of the propylene tower. Construct a penalty function according to formula (1) to impose penalties on solutions that deviate from the operating parameter constraints of the propylene distillation tower. After updating the current propylene production and energy consumption optimization objective function values, perform non-dominated sorting on the wolf pack individual solutions to select the non-dominated solution of the current population. Archive The population individuals are compared and the dominated individual solutions are deleted to complete the pairing. Archive Population update.

[0015] S6.5: Judgment ArchiveWhether the population has reached the upper limit N W If the limit is exceeded, calculate the current Archive The crowding value distribution of the optimization objective function value (propylene tower output, propylene tower total energy consumption) of each individual in the group, and the overcrowded individual solutions are eliminated. The specific operations are as follows: ① Calculate Archive The two optimization objective function values ​​of propylene production and total energy consumption of propylene tower for each individual in the population are calculated, and the maximum and minimum values ​​of each objective function are found and recorded as and ② Calculate the crowding distance of each individual : In the formula, , Represented as individual The adjacent two ③Each individual in the population has a crowding distance corresponding to the objective function. The population is above the upper limit N W When the algorithm randomly deletes individuals with low crowding, and retains individuals with high crowding to maintain The number of individuals; S7. Determine whether the current maximum number of iterations has reached the upper limit t_max , if yes, then output the non-dominated individuals in the population as the multi-objective optimization result of the propylene tower, otherwise return to S4 to continue the iterative calculation.

[0016] Fig.13 The optimization results show that under the same device energy consumption, the increase in production can be achieved by adjusting the operating parameter combination of the propylene tower. That is, under the same distillation tower production energy consumption, the propylene production can be maximized by changing the parameter combination of the propylene tower feed temperature and reflux flow. Figure 4 The five-pointed star in the figure represents the actual operating condition of the propylene tower (output 10788.33 kg / hr, energy consumption 34.7 M*KCAL / HR). By comparing the point A which is close to the actual operating condition output through the arrow-enlarged area chart, its propylene output is 10786.56 kg / hr, and its energy consumption is 33.4 M*KCAL / HR. The optimization of point A saves 1.3 KCAL / HR. Compared with point B which is close to the actual energy consumption, its energy consumption is 34.72 M*KCAL / HR, and its propylene output is 10840.87 kg / hr. The optimization of point B increases the output by 52.54 kg / hr.

[0017] Fig.14 The parameter combination optimization scheme for three groups of feed temperature and reflux flow is given. In order to verify the accuracy of the model prediction, Fig.14The three Pareto optimal solutions are brought into the propylene distillation unit model of PRO / II, and the simulation results are listed in Fig.14 After comparing the simulated values ​​and predicted values ​​in the table, it is found that their relative deviations are all less than 5%, which shows that the proxy model has a better prediction effect on the target variable and the optimized process conditions are reliable.

[0018] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A multi-objective energy-saving optimization method for a propylene distillation tower based on state simulation and improved wolf pack algorithm, characterized in that: The following steps are involved: S1. Based on the equilibrium-level steady-state model of the distillation process, the Peng-Robinson (PR) equation combined with the Inside-out convergence algorithm was used to simulate the production state of the propylene distillation tower based on the PRO / II process simulation platform, and a state simulation model of the propylene distillation tower was established; S2. Based on the state simulation model of the propylene distillation tower, the sensitivity analysis of single parameters and combined parameters of the operating parameters of the propylene distillation tower: tower top pressure, raw material feed temperature, tower top reflux rate is performed to establish a data relationship table between the propylene production of the propylene distillation tower and the energy consumption of the propylene tower and the raw material feed temperature and tower top reflux rate; S3. Use multiple regression analysis and least squares method to identify model parameters of the propylene distillation tower data relationship table, establish single-objective optimization functions of propylene tower propylene production and propylene tower energy consumption, and establish a multi-objective energy-saving optimization model for the propylene tower in combination with the actual operating parameter constraints of the propylene distillation tower; S4. According to formula (1), let the penalty factor be r k , construct the penalty function term Establish propylene production penalty function Φ1 and propylene tower energy consumption penalty function Φ2, and transform the constrained propylene tower multi-objective optimization into unconstrained multi-objective optimization; In the formula, u is the number of initial iterations, n To terminate the iterations, is the objective function to be optimized, is the inequality constraint, express When the larger value of When the constraints are met, =0, penalty term =0, when When the constraints are not met, 0, penalty item 0, penalty factor r k+1 =c r k , k=k+ 1. As the iteration of the algorithm gradually increases, the penalty for the optimization objective function increases; S5. In order to improve the local search ability of the optimization algorithm, speed up the convergence speed, and avoid the algorithm falling into the local optimum, a dynamic search factor is added θ Improve the ability of the wolf pack algorithm to search for the optimal solution, and calculate according to formula (2) θ The current value of In the formula, , is the dynamic search factor, t , t_max The algorithm is the current number of iterations and the maximum number of iterations; S6. Execute the improved multi-objective wolf pack algorithm to solve the unconstrained multi-objective optimization function of the propylene distillation tower, including the following sub-steps: Step S6.1: After calculating the multi-objective optimization function value of the propylene tower (maximizing propylene production and minimizing propylene tower energy consumption) of each individual solution in the wolf pack algorithm, perform non-dominated sorting on the current optimal solution and sort it according to the population size. Upper limit update Archive Optimal solution population; Step S6.2: Calculate parameters α and algorithm swing factor A , C The value of the parameter is shown in formula (3): In the formula, α The control parameter takes values ​​in the range [0,2]. r 1. r 2 is a random number in the range [0,1], t , t_max is the current number of iterations and the maximum number of iterations; Step S6.3: Use roulette wheel selection method to select Archive Select the optimal solution of the two optimization objective functions of the current propylene tower propylene production and the total energy consumption of the propylene distillation tower from the optimal solution population α , optimal solution β , suboptimal solution δ , and update the optimal orientation of each wolf pack algorithm in the population according to formulas (2), (4), and (5), as shown in formulas (4) and (5): In the formula, The updated search direction of the wolf pack algorithm; Step S6.4: Calculate the objective function value of the new iterative search direction of the wolf pack individuals, that is, the current propylene production and energy consumption of the propylene tower. Construct a penalty function according to formula (1) to impose penalties on solutions that deviate from the operating parameter constraints of the propylene distillation tower. After updating the current propylene production and energy consumption optimization objective function values, perform non-dominated sorting on the wolf pack individual solutions to select the non-dominated solution of the current population. Archive The population individuals are compared and the dominated individual solutions are deleted to complete the pairing. Archive Population update.

2. Step S6.5: Determination Archive Whether the population has reached the upper limit N W If the limit is exceeded, calculate the current Archive The crowding value distribution of the optimization objective function value (propylene tower output, propylene tower total energy consumption) of each individual in the group is calculated, and the individual solutions with too dense crowding are eliminated. 3.S7. Determine whether the current maximum number of iterations has reached the upper limit t_max , if yes, then output the non-dominated individuals in the population as the multi-objective optimization result of the propylene tower, otherwise return to S4 to continue the iterative calculation.

Images