A Multi-Agent Automatic Control and Simulation Method for Cluster Distillation Columns

Abstract

This invention discloses a multi-agent automatic control and simulation calculation method for a cluster distillation column. The method includes: columnar distillation units within the cluster distillation column are formed by dividing the column's internal space into uniformly distributed units that are independent of each other, with no mass transfer between any two units; simulating the separation performance of individual units; coupling multiple units to form a cluster distillation column and designing a multi-loop process control system for the device; designing an intelligent agent unit controller to address the multi-coupling characteristics of the cluster distillation column, achieving uniform fluid distribution and state consistency among the units; performing dynamic simulation calculations of the multi-loop process control system, optimizing controller parameters, and testing anti-interference performance. By using simulation calculation technology to analyze key operating parameters of the cluster distillation column device, determining optimal distillation conditions, and designing a multi-agent automatic control system, this method provides a technical foundation for the industrialization of cluster distillation columns.

Description

Technical Field

[0001] This invention relates to the field of chemical technology application technology, specifically to a multi-agent automatic control and simulation calculation method for a cluster distillation column. Background Technology

[0002] Currently, distillation accounts for over 70% of chemical separation processes. Industrial distillation columns are large in size and numerous, with long residence times and large storage volumes of materials within them. Especially for difficult-to-separate systems with similar relative volatility and azeotropic systems, the theoretical number of plates often reaches hundreds to achieve separation efficiency. This necessitates constructing distillation columns approaching 100 meters in height, significantly increasing the associated safety risks. Developing efficient packing materials, novel trays, and internal components can theoretically improve separation efficiency, but current industrial distillation units lack significant improvement measures.

[0003] Invention Patent (A Cluster Distillation Column, CN202110144285.6). This cluster distillation column uses centimeter-scale distillation units filled with high-efficiency packing materials. By integrating distillation clusters into an array, a containerized distillation unit is constructed, significantly reducing the liquid holdup within each unit and greatly mitigating safety risks. This cluster distillation equipment enables the industrial separation of difficult-to-separate materials with relative volatility close to 1, eliminating the need for distillation columns tens of meters high and greatly reducing equipment investment costs.

[0004] However, during the separation process, the operating efficiency, target quality of the mixture being processed, economic efficiency and safety of the production operation of the cluster distillation column are all affected by many factors. In addition, the complexity of the coupling between distillation units leads to a high degree of uncertainty and complexity in the cluster distillation column. Summary of the Invention

[0005] In view of this, the present invention provides a multi-agent automatic control and simulation calculation method for cluster distillation columns, develops multi-agent automatic control technology, and achieves integrated and precise control of the "pile / cluster" array, thereby improving equipment reliability. This method provides a technical foundation for achieving sustainable and efficient cluster distillation column production.

[0006] Technical solution

[0007] To achieve the above objectives, the present invention employs the following technical solution:

[0008] This invention provides a multi-agent automatic control and simulation calculation method for a cluster distillation column, comprising:

[0009] (1) The separation performance of a single distillation unit in a cluster distillation column was simulated and calculated, as follows:

[0010] 1) Determining the physical properties of the separation system and the simulation modules: The physical properties of the system were determined using chemical process simulation software based on the characteristics of the reactive distillation process. The selected modules in the simulation included: Heater module for heat exchangers, RadFrac module for columns, Mixer module for mixers, FSplit module for distributors, Flash2 module for flash tanks, and Pump module for pressurization pumps.

[0011] (2) Determine the optimal steady-state process parameters: Based on the system characteristics and actual process parameters of the reactive distillation process, establish a steady-state model of the distillation unit using chemical process simulation software. On the basis of the established steady-state model, determine the optimal steady-state process structure parameters and operating parameters of the distillation unit.

[0012] (3) Determine the control objectives and control degrees of freedom of the control system: Based on the steady-state design, and considering the temperature, gas-liquid phase flow rate and gas-liquid composition distribution of the distillation unit, the product quality and target product selectivity should be reasonably determined as the control objectives of the control system design according to the process requirements.

[0013] (2) Multiple distillation units are coupled together to form a cluster distillation column, specifically:

[0014] Ten distillation units are connected in parallel to form a cluster distillation column. The condenser and reboiler are separately located outside the unit, and the material is uniformly fed into each distillation unit. The liquid material at the bottom of the cluster distillation column is vaporized in the reboiler and enters the distillation unit through the vapor inlet, forming rising vapor that is evenly distributed into each distillation unit. The rising vapor leaves the cluster distillation column through the vapor outlet and enters the subsequent condenser to form condensate. A portion of the condensate is evenly distributed into each distillation unit in the rectification section through a reflux device. A portion of the material from the reboiler and condenser leaves the cluster distillation column at an appropriate concentration and flow rate, completing the entire distillation separation process.

[0015] (3) Design of a multi-loop process control system for the cluster distillation column unit, specifically:

[0016] 1) Feed flow is controlled by flow control. The controller operates in a reactive mode, controlling or adjusting the feed flow by adjusting its setpoint.

[0017] 2) Control the pressure at the top of the distillation column by adjusting the heat transfer of the top condenser.

[0018] 3) A composition-flow cascade control scheme for cluster distillation columns is used to regulate the reflux flow rate of the overhead stream. A component controller is added to the overhead product stream of the cluster distillation column to control and maintain the concentration of the overhead product. A cascade control loop is formed by adjusting the setpoint of the flow controller. The component controller is the primary controller, and the flow controller is the secondary controller.

[0019] 4) The temperature of the sensitive trays in the distillation unit is controlled by adjusting the heat load of the reboiler in the column bottom.

[0020] 5) In the cluster distillation column unit, the liquid level of the reboiler in the column bottom is controlled by adjusting the feed flow rate from the column bottom.

[0021] (4) A smart agent unit controller is designed to address the multi-coupling characteristics of cluster distillation columns, achieving consistent fluid distribution within the unit. Specifically:

[0022] The design is based on the relative states between intelligent agents to design a control input feed flow rate u. i (t) enables consistency in the MAN implementation of a linear second-order dynamical description of a single agent. The definition of consistency is as follows:

[0023] If for any i, j = 1, 2, ..., N and any initial state, we have

[0024]

[0025] MAN achieves consistency. To achieve consistency, the consistency control input of MAN is designed as follows:

[0026]

[0027] Where K is the gain of the controller to be designed.

[0028] Based on the above, the adjacency matrix is ​​defined under the Markov switching topology. And there are

[0029] Therefore, the consistency control flow input can also be designed as

[0030]

[0031] Where, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij Let K represent the adjacency matrix between the i-th and j-th agents. r(t) The control gain to be designed is ρ(t) = ρ ∈ S = {1, 2, ..., N}, which is the switching signal that determines the communication topology pattern. r(t) is a hidden Markov chain that depends on a Markov process, and g(x) is the control gain to be designed. i (t) represents the controller disturbance caused when the system switches topologies, and it is assumed to be bounded.

[0032] (5) Optimize the controller parameters for the cluster distillation column unit, specifically as follows:

[0033] The control loop is designed using a proportional-integral-derivative (PID) controller. Dead time is introduced into the temperature and composition control loops. In dynamic simulation, a delay module is used to characterize process delay, and a first-order module is used to increase the dead time of the process and equipment. Relay feedback tests are performed sequentially on the temperature and composition controllers using the precise Tyreus-Luyben rule for parameter tuning and refinement. The Tyreus-Luyben tuning formula is...

[0034]

[0035] (6) Testing the anti-interference performance of the cluster distillation column unit, specifically:

[0036] For the design of a multi-loop automatic control system for a cluster distillation column, feed flow disturbances and feed composition disturbances are introduced in dynamic simulation to analyze the dynamic response of the system. If the dynamic response meets the requirements of the control system, the design scheme is effective; otherwise, the controller parameters are readjusted and the loops with poor control performance in the control loop are modified until the dynamic response meets the system requirements.

[0037] Beneficial effects

[0038] The technical solution provided by this invention has the following advantages compared with known public technologies:

[0039] Using simulation technology, a series of parameters such as feed state, material feed rate, distillation temperature, and pressure of each distillation unit in a cluster distillation column are simulated and analyzed to determine the optimal distillation conditions. The design and dynamic simulation calculation of a multi-loop process control system for the cluster distillation column provide a technical foundation for industrial cluster distillation column devices, guiding practice through theory. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the process simulation of a cluster distillation column according to the present invention;

[0041] Figure 2 This is a schematic diagram of a multi-agent automatic control system for a cluster distillation column according to the present invention;

[0042] Figure 3 To implement Case 1, the dynamic response diagram of the system after adding feed flow rate and feed disturbance is shown ((1) o-chlorotoluene content under feed flow rate disturbance; (2) o-chlorotoluene content under feed flow rate disturbance).

[0043] Figure 4 To implement Case 2, the dynamic response diagram of the system after adding feed flow rate and feed disturbance is shown ((1) content of o-chlorobenzyl chloride under feed flow rate disturbance; (2) content of o-chlorobenzyl chloride under feed flow rate disturbance).

[0044] Figure 5 To implement Case 3, the dynamic response diagram of the system after adding feed flow rate and feed disturbance is shown ((1) content of benzyl chloride under feed flow rate disturbance; (2) content of benzyl chloride under feed flow rate disturbance). Detailed Implementation

[0045] Implementation Case 1: Simulation Case of o- / p-chlorotoluene separation process in a cluster distillation column unit

[0046] Step 1: Use simulation software to determine the method for separating the physical properties of the system and the simulation unit modules.

[0047] The chemical process simulation software (Aspen in this embodiment, but APEX and other software can also be used in other embodiments) is used to determine the system composition based on the characteristics of the reactive distillation process using the PR (Peng-Robinson) equation of state. The selected unit modules in the simulation include: Heater module for heat exchangers, RadFrac module for columns, Mixer module for mixers, FSplit module for distributors, Flash2 module for flash tanks, and Pump module for pressurization pumps.

[0048] Step 2: Simulate the distillation separation process using a single distillation unit as the object:

[0049] Based on the gas wave equilibrium constant, boiling point, relative volatility, and other system characteristics and actual process parameters of the distillation process system, a steady-state model of the ortho / p-chlorotoluene separation process is established using chemical process simulation software, such as... Figure 1 As shown in Table 1, (steady-state mechanism models of different distillation columns are built according to different separation systems). Based on the established steady-state model of the process flow, the optimal steady-state process structure parameters and operating parameters of the multi-unit reactive distillation unit are determined (the established steady-state model is simulated and tested in the software, and the optimal structure parameters and operating parameters are determined based on the test results).

[0050] Step 3: Determine the control objective and degrees of freedom of the ortho / para-chlorotoluene separation control system.

[0051] Based on steady-state conditions, considering the tray temperature, gas-liquid flow rate, and gas-liquid composition distribution within the column, and according to process requirements, product quality and target product selectivity are rationally determined as the control objectives for the control system design. Based on the steady-state process simulation, the control degrees of freedom (i.e., controlled variables) of the control system are determined. Regarding production indicators, when designing the control system for the o- / p-chlorotoluene separation process in a cluster distillation column, the requirement is that the o-chlorotoluene content (MoleFraction) be stable at around 99.5%. In the o- / p-chlorotoluene separation process of the cluster distillation column, several controlled variables are used to control the entire process. These are the pressure, temperature, reflux flow rate, feed rate, subcooling temperature, and purity of the extracted o- / p-chlorotoluene in each distillation unit.

[0052] Table 1

[0053]

[0054]

[0055] Step 4: Couple multiple distillation units together to form a cluster distillation column;

[0056] Based on the simulation of a single distillation unit, ten distillation units (RadFrac1-RadFrac6) were connected in parallel to construct a cluster distillation column simulation device. The feed (a mixture of ortho / para-chlorotoluene) was uniformly fed into each distillation unit. The liquid feed at the bottom of the cluster distillation column was vaporized in the reboiler (Heater2) and entered the distillation unit through the vapor inlet, forming rising vapor that was uniformly distributed into each distillation unit. The rising vapor left the cluster distillation column through the vapor outlet and entered the subsequent condenser (Heater1) to form condensate. A portion of the condensate was then uniformly distributed into each distillation unit in the rectification section through a reflux device. A portion of the feed from the reboiler and condenser left the cluster distillation column device at a concentration and flow rate meeting the separation requirements, completing the entire ortho / para-chlorotoluene distillation separation process.

[0057] After the above simulation process was completed, the simulation results of the key logistics are shown in Table 2:

[0058] Table 2

[0059] Cluster distillation column operating parameters numerical values Simulated pressure (mmHg) 92 Temperature of materials transported at the top of the tower (°C) 92.28 Top distillate (kg / h) 5.32 molar reflux ratio 35.4 Retort temperature (°C) 115.06 Distillate output from the reboiler (kg / h) 5.32 o-chlorotoluene content in the top discharge of the tower 99.5 p-chlorotoluene content in the bottom feed of the tower 99.5

[0060] As can be seen from the table above, the o-chlorotoluene content in the top feed of the tower is 99.5%, which well meets the requirements of the o / p-chlorotoluene separation process.

[0061] Step 5: Design a multi-loop process control system for the cluster distillation column unit for ortho / para-chlorotoluene separation. Select temperature, pressure, and liquid level process variables closely related to the production process for the control loop design, such as... Figure 2 As shown.

[0062] 1) Feed flow is controlled by flow control. The controller operates in a reactive mode, controlling or adjusting the feed flow by adjusting its setpoint.

[0063] 2) Control the pressure at the top of the distillation column by adjusting the heat transfer of the top condenser.

[0064] 3) A composition-flow cascade control scheme is used to regulate the reflux flow rate of the distillation column. A component controller is added to the top product stream of the distillation column to control and maintain the concentration of o-chlorotoluene. The cascade control loop is formed by adjusting the setpoint of the flow controller. The component controller is the primary controller, and the flow controller is the secondary controller.

[0065] 4) The temperature of the sensitive trays in the distillation unit is controlled by adjusting the heat load of the reboiler in the column bottom.

[0066] 5) The liquid level in the reboiler of the stacking device is controlled by adjusting the reboiler discharge flow rate.

[0067] Step 6: Design an intelligent agent unit controller for the multi-coupling characteristics of the cluster distillation column for ortho / para-chlorotoluene separation to achieve consistent fluid distribution within the unit, specifically:

[0068] Design a flow input u based on the relative states between agents. i (t), enabling consistency in the MAN described by the linear second-order dynamics of a single agent. Therefore, the consistency control flow input can be designed as

[0069]

[0070] Where, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij Let K represent the adjacency matrix between the i-th and j-th agents. r(t) The control gain to be designed The switching signal determines the communication topology pattern; r(t) is a hidden Markov chain that depends on a Markov process; g(x) i (t) represents the controller disturbance caused when the system switches topologies, and it is assumed to be bounded.

[0071] Step 7: Optimize the controller parameters for the cluster distillation column unit for separating ortho / para-chlorotoluene, specifically as follows;

[0072] 1) Set the proportional gain Kc of all level controllers to 2 and the integral time to 9999 minutes.

[0073] 2) Set the proportional gain Kc of all pressure controllers to 5 and the integral time to 20 minutes.

[0074] 3) A flow controller with valve opening as the manipulated variable, with proportional gain Kc set to 0.5 and integral time of 0.3 minutes.

[0075] Dead time is introduced into the temperature and component control loops: 1 minute and 2 minutes are set in the temperature control loops of the distillation column and reactor, respectively, and 3 minutes in the component control loop. In the dynamic simulation, delay modules are used to characterize process delays, and first-order modules are used to increase the dead time of the process and equipment. Accurate Tyreus-Luyben rules are used to perform relay feedback tests on the temperature and component controllers sequentially for parameter tuning and refinement.

[0076] Taking the temperature control loop in the RadFrac1 distillation unit as an example, it is first set as a closed loop with a 5% disturbance amplitude. Then, relay feedback testing is performed to obtain the dynamic response and oscillation curve of the control loop. Finally, the Tyreus-Luyben rule is applied to calculate the limiting period Pu = 0.33856 and the limiting gain K of the oscillation curve. n =6 minutes. The test will end after which... The proportional gain Kc = 0.1058 and the integration time τ1 = 13.2 minutes for the corresponding loop were calculated.

[0077] Step 8: Disturbance operation, analyze control system performance

[0078] Introducing disturbances, analyzing the system's dynamic response performance indicators, evaluating the effectiveness of the design scheme, and modifying the control loop as needed to ensure the system meets control requirements. The system response under disturbances in feed flow rate and feed composition is as follows: Figure 3 The figure shows the system's settling time. Performance testing demonstrates that this control system has good control performance, can effectively overcome disturbances in the production process, and meets the requirements of industrial production control.

[0079] Implementation Case 2: Simulation Case of the Separation Process of o-chlorobenzyl chloride and o-chlorobenzylidene dichloro in a Cluster Distillation Column

[0080] Step 1: Use simulation software to determine the method for separating the physical properties of the system and the simulation unit modules.

[0081] The system characteristics of the reactive distillation process were determined using the PR (Peng-Robinson) equation of state method in chemical process simulation software. The selected unit modules in the simulation included: Heater module for heat exchangers, RadFrac module for columns, Mixer module for mixers, FSplit module for distributors, Flash2 module for flash tanks, and Pump module for pressurization pumps.

[0082] Step 2: Simulate the distillation separation process using a single distillation unit as the object:

[0083] Based on the gas wave equilibrium constant, boiling point, relative volatility, and other system characteristics and actual process parameters of the distillation process system, a steady-state model of the separation process of o-chlorobenzyl chloride and o-chlorobenzylidene dichloride was established using chemical process simulation software, such as... Figure 1 As shown in Table 3, based on the established steady-state model of the process flow, the optimal steady-state process structure parameters and operating parameters of the multi-unit reactive distillation unit are determined.

[0084] Step 3: Determine the control objective and degrees of freedom of the o-chlorobenzyl chloride and o-chlorobenzyl dichloro dichloride separation control system.

[0085] Based on steady-state conditions, considering the tray temperature, gas-liquid flow rate, and gas-liquid composition distribution within the column, and according to process requirements, product quality and target product selectivity are rationally determined as the control objectives for the control system design. Based on the steady-state process simulation, the control degrees of freedom (i.e., controlled variables) of the control system are determined. Regarding production indicators, when designing the control system for the separation process of o-chlorobenzyl chloride and o-chlorobenzylide dichloride in a cluster distillation column, the requirement is that the mole fraction of benzyl chloride benzyl chloride be kept stable at around 95%. In the separation process of o-chlorobenzyl chloride and o-chlorobenzylide dichloride in the cluster distillation column, several controlled variables are used to control the entire process. These are the pressure, temperature, reflux flow rate, feed rate, subcooling temperature, and purity of the extracted benzyl chloride in each distillation unit.

[0086] Table 3

[0087]

[0088] Step 4: Couple multiple distillation units together to form a cluster distillation column;

[0089] Based on the simulation of a single distillation unit, ten distillation units (RadFrac1-RadFrac6) were connected in parallel to construct a cluster distillation column simulation device. The feed (a mixture of o-chlorobenzyl chloride and o-chlorobenzyl dichlorodichloro) was uniformly fed into each distillation unit. The liquid feed from the bottom of the cluster distillation column was vaporized in the reboiler (Heater2) and entered the distillation unit through the vapor inlet, forming rising vapor that was uniformly distributed into each distillation unit. The rising vapor left the cluster distillation column through the vapor outlet and entered the subsequent condenser (Heater1) to form condensate. A portion of the condensate was then uniformly distributed into each distillation unit in the rectification section through a reflux device. A portion of the feed from the reboiler and condenser left the cluster distillation column device at a concentration and flow rate meeting the separation requirements, completing the entire o / p-chlorotoluene distillation separation process.

[0090] After the above simulation process was completed, the simulation results of the key logistics were shown in Table 4:

[0091] Table 4

[0092] Cluster distillation column operating parameters numerical values Simulated pressure (mmHg) 97 Temperature of materials transported at the top of the tower (°C) 101.5 Top distillate (kg / h) 2.87 molar reflux ratio 3.1 Retort temperature (°C) 113 o-chlorobenzyl chloride content in the top discharge of the tower 95

[0093] As can be seen from the table above, the content of o-chlorobenzyl chloride in the top discharge of the tower is 95%, which well meets the requirements of the o-chlorobenzyl chloride and o-chlorobenzyl dichlorodichloro separation process.

[0094] Step 5: Design a multi-loop process control system for the cluster distillation column unit separating o-chlorobenzyl chloride and o-chlorobenzyl dichlorodichloro. Select temperature, pressure, and liquid level process variables closely related to the production process for the control loop design, such as... Figure 2 As shown.

[0095] 1) Feed flow is controlled by flow control. The controller operates in a reactive mode, controlling or adjusting the feed flow by adjusting its setpoint.

[0096] 2) Control the pressure at the top of the distillation column by adjusting the heat transfer of the top condenser.

[0097] 3) A composition-flow cascade control scheme is used to regulate the reflux flow rate of the distillation column. A component controller is added to the top product stream of the distillation column to control and maintain the concentration of o-chlorotoluene. The cascade control loop is formed by adjusting the setpoint of the flow controller. The component controller is the primary controller, and the flow controller is the secondary controller.

[0098] 4) The temperature of the sensitive trays in the distillation unit is controlled by adjusting the heat load of the reboiler in the column bottom.

[0099] 5) The liquid level in the reboiler of the stacking device is controlled by adjusting the reboiler discharge flow rate.

[0100] Step 6: Design an intelligent agent unit controller for the multi-coupling characteristics of the cluster distillation column separating o-chlorobenzyl chloride and o-chlorobenzyl dichloride, to achieve consistent fluid distribution within the unit, specifically:

[0101] Design a flow input u based on the relative states between agents. i (t), enabling consistency in the MAN described by the linear second-order dynamics of a single agent. Therefore, the consistency control flow input can be designed as

[0102]

[0103] Where, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij Let K represent the adjacency matrix between the i-th and j-th agents. r(t) The control gain to be designed The switching signal determines the communication topology pattern; r(t) is a hidden Markov chain that depends on a Markov process; g(x) i (t) represents the controller disturbance caused when the system switches topologies, and it is assumed to be bounded.

[0104] Step 7: Optimize the controller parameters of the cluster distillation column unit for separating o-chlorobenzyl chloride and o-chlorobenzyl dichlorodichloro, specifically as follows;

[0105] 1) Set the proportional gain Kc of all level controllers to 2 and the integral time to 9999 minutes.

[0106] 2) Set the proportional gain Kc of all pressure controllers to 5 and the integral time to 20 minutes.

[0107] 3) A flow controller with valve opening as the manipulated variable, with proportional gain Kc set to 0.5 and integral time of 0.3 minutes.

[0108] Dead time is introduced into the temperature and component control loops: 1 minute and 2 minutes are set in the temperature control loops of the distillation column and reactor, respectively, and 3 minutes in the component control loop. In the dynamic simulation, delay modules are used to characterize process delays, and first-order modules are used to increase the dead time of the process and equipment. Accurate Tyreus-Luyben rules are used to perform relay feedback tests on the temperature and component controllers sequentially for parameter tuning and refinement.

[0109] Taking the temperature control loop in the RadFrac1 distillation unit as an example, it is first set as a closed loop with a 5% disturbance amplitude. Then, relay feedback testing is performed to obtain the dynamic response and oscillation curve of the control loop. Finally, the limiting period P of the oscillation curve is calculated using the Tyreus-Luyben rule. u =0.4458 and the limiting gain K u =5 minutes. The test will end after which... The proportional gain Kc = 0.258 and the integration time τ1 = 9.2 minutes for the corresponding loop were calculated.

[0110] Step 8: Disturbance operation, analyze control system performance

[0111] Introducing disturbances, analyzing the system's dynamic response performance indicators, evaluating the effectiveness of the design scheme, and modifying the control loop as needed to ensure the system meets control requirements. The system response under disturbances in feed flow rate and feed composition is as follows: Figure 4 The figure shows the system's settling time. Performance testing demonstrates that this control system has good control performance, can effectively overcome disturbances in the production process, and meets the requirements of industrial production control.

[0112] Implementation Case 3: Simulation Case of the Separation Process of p-chlorobenzyl chloride / p-chlorobenzyl dichloroide in a Cluster Distillation Column

[0113] Step 1: Use simulation software to determine the method for separating the physical properties of the system and the simulation unit modules.

[0114] The system characteristics of the reactive distillation process were determined using the PR (Peng-Robinson) equation of state method in chemical process simulation software. The selected unit modules in the simulation included: Heater module for heat exchangers, RadFrac module for columns, Mixer module for mixers, FSplit module for distributors, Flash2 module for flash tanks, and Pump module for pressurization pumps.

[0115] Step 2: Simulate the distillation separation process using a single distillation unit as the object:

[0116] Based on the gas wave equilibrium constant, boiling point, relative volatility, and other system characteristics and actual process parameters of the distillation process system, a steady-state model of the separation process of p-chlorobenzyl chloride / p-chlorobenzylidene dichlorodichloro was established using chemical process simulation software, such as... Figure 1 As shown in Table 5, based on the established steady-state model of the process flow, the optimal steady-state process structure parameters and operating parameters of the multi-unit reactive distillation unit are determined.

[0117] Step 3: Determine the control objective and degrees of freedom of the p-chlorobenzyl chloride / p-chlorobenzyl dichloro dichloro separation control system.

[0118] Based on steady-state conditions, considering the tray temperature, gas-liquid flow rate, and gas-liquid composition distribution within the column, and according to process requirements, product quality and target product selectivity are rationally determined as the control objectives for the control system design. Based on the steady-state process simulation, the control degrees of freedom (i.e., controlled variables) of the control system are determined. Regarding production indicators, when designing the control system for the p-chlorobenzyl chloride / p-chlorobenzylide dichloro separation process in a cluster distillation column, the requirement is that the p-chlorobenzyl chloride content (MoleFraction) be stable at around 95%. In the p-chlorobenzyl chloride / p-chlorobenzylide dichloro separation process of the cluster distillation column, several controlled variables are used to control the entire process. These are the pressure, temperature, reflux flow rate, feed rate, subcooling temperature, and purity of the extracted p-chlorobenzyl chloride in each distillation unit.

[0119] Table 5

[0120]

[0121] Step 4: Couple multiple distillation units together to form a cluster distillation column;

[0122] Based on the simulation of a single distillation unit, ten distillation units (RadFrac1-RadFrac10) were connected in parallel to construct a cluster distillation column simulation device. The feed (a mixture of p-chlorobenzyl chloride and p-chlorobenzyl dichlorodichloro) was uniformly fed into each distillation unit. The liquid feed from the bottom of the cluster distillation column was vaporized in the reboiler (Heater2) and entered the distillation unit through the vapor inlet, forming rising vapor that was uniformly distributed into each distillation unit. The rising vapor left the cluster distillation column through the vapor outlet and entered the subsequent condenser (Heater1) to form condensate. A portion of the condensate was then uniformly distributed into each distillation unit in the rectification section through a reflux device. A portion of the feed from the reboiler and condenser left the cluster distillation column device at a concentration and flow rate meeting the separation requirements, completing the entire ortho / p-chlorotoluene distillation separation process.

[0123] After the above simulation process was completed, the simulation results of the key logistics are shown in Table 6:

[0124] Table 6

[0125] Cluster distillation column operating parameters numerical values Simulated pressure (mmHg) 97 Temperature of materials transported at the top of the tower (°C) 110 Top distillate (kg / h) 3.62 molar reflux ratio 1.9 Retort temperature (°C) 126.06 The content of benzyl chloride in the top discharge of the tower 95

[0126] As can be seen from the table above, the content of benzyl chloride in the top discharge of the tower is 95%, which well meets the requirements of the benzyl chloride / benzyl chloride dichlorodichloro separation process.

[0127] Step 5: Design a multi-loop process control system for the cluster distillation column unit for the separation of benzyl chloride / benzyl chloride dichlorodi ... Figure 2 As shown.

[0128] 1) Feed flow is controlled by flow control. The controller operates in a reactive mode, controlling or adjusting the feed flow by adjusting its setpoint.

[0129] 2) Control the pressure at the top of the distillation column by adjusting the heat transfer of the top condenser.

[0130] 3) A composition-flow cascade control scheme is used to regulate the reflux flow rate of the distillation column. A component controller is added to the top product stream of the distillation column to control and maintain the concentration of o-chlorotoluene. The cascade control loop is formed by adjusting the setpoint of the flow controller. The component controller is the primary controller, and the flow controller is the secondary controller.

[0131] 4) The temperature of the sensitive trays in the distillation unit is controlled by adjusting the heat load of the reboiler in the column bottom.

[0132] 5) The liquid level in the reboiler of the stacking device is controlled by adjusting the reboiler discharge flow rate.

[0133] Step 6: Design an intelligent agent unit controller for the multi-coupling characteristics of cluster distillation columns separating p-chlorobenzyl / p-chlorobenzyl dichlorodichloro, to achieve consistent fluid distribution within the unit, specifically:

[0134] Design a flow input u based on the relative states between agents. i (t), enabling consistency in the MAN described by the linear second-order dynamics of a single agent. Therefore, the consistency control flow input can be designed as

[0135]

[0136] Where, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij Let K represent the adjacency matrix between the i-th and j-th agents. r(t) The control gain to be designed The switching signal determines the communication topology pattern; r(t) is a hidden Markov chain that depends on a Markov process; g(x) i (t) represents the controller disturbance caused when the system switches topologies, and it is assumed to be bounded.

[0137] Step 7: Optimize the controller parameters of the cluster distillation column unit for the separation of p-chlorobenzyl chloride / p-chlorobenzyl dichlorodichloro, specifically as follows;

[0138] 1) Set the proportional gain Kc of all level controllers to 2 and the integral time to 9999 minutes.

[0139] 2) Set the proportional gain Kc of all pressure controllers to 5 and the integral time to 20 minutes.

[0140] 3) A flow controller with valve opening as the manipulated variable, with proportional gain Kc set to 0.5 and integral time of 0.3 minutes.

[0141] Dead time is introduced into the temperature and component control loops: 1 minute and 2 minutes are set in the temperature control loops of the distillation column and reactor, respectively, and 3 minutes in the component control loop. In the dynamic simulation, delay modules are used to characterize process delays, and first-order modules are used to increase the dead time of the process and equipment. Accurate Tyreus-Luyben rules are used to perform relay feedback tests on the temperature and component controllers sequentially for parameter tuning and refinement.

[0142] Taking the temperature control loop in the RadFrac1 distillation unit as an example, it is first set as a closed loop with a 5% disturbance amplitude. Then, relay feedback testing is performed to obtain the dynamic response and oscillation curve of the control loop. Finally, the Tyreus-Luyben rule is applied to calculate the limiting period Pu = 0.4786 and the limiting gain K of the oscillation curve. u =8 minutes. The test will end after which... P U The proportional gain Kc = 0.2563 and the integration time τ1 = 7.5 minutes for the corresponding loop were calculated.

[0143] Step 8: Disturbance operation, analyze control system performance

[0144] Introducing disturbances, analyzing the system's dynamic response performance indicators, evaluating the effectiveness of the design scheme, and modifying the control loop as needed to ensure the system meets control requirements. The system response under disturbances in feed flow rate and feed composition is as follows: Figure 5 The figure shows the system's settling time. Performance testing demonstrates that this control system has good control performance, can effectively overcome disturbances in the production process, and meets the requirements of industrial production control.

Claims

1. A multi-agent automatic control and simulation calculation method for a cluster distillation column, wherein the columnar distillation units within the cluster distillation column are formed by dividing the internal space of the column, the distillation units are uniformly distributed, and the distillation units are independent of each other, with no mass transfer between any adjacent distillation units, characterized in that... The separation performance of a single distillation unit is simulated and calculated. Multiple distillation units are coupled and combined to form a cluster distillation column. A multi-loop process control system for the cluster distillation column is designed, and temperature, pressure, and liquid level are selected as the controlled variables. A smart agent unit controller is designed for the multi-coupling characteristics of the cluster distillation column to achieve uniform fluid distribution and state consistency among the units in the cluster distillation column. Dynamic simulation calculations of the multi-loop process control system are performed, and the controller parameters are optimized and the anti-interference performance is tested. A smart agent unit controller is designed to address the multi-coupling characteristics of cluster distillation columns, achieving consistent fluid distribution within the unit. Specifically: The design goal is to design a control input based on the relative states between agents. This enables consistency in the MAN (Manifestational Architecture) described by a single agent's linear second-order dynamics. Therefore, the consistency control flow input is designed as follows: in, Indicates neighbor status information, Indicates the first State information of each agent. Indicates the first and the Adjacency matrix of agents The control gain to be designed The switching signal determines the mode of the communication topology. It is a hidden Markov chain that depends on Markov processes. The controller disturbance caused by system topology switching is assumed to be bounded. The controller parameters for the cluster distillation column unit were optimized as follows: A proportional-integral-derivative (PID) controller is applied to design the control loop, introducing dead time into the temperature and composition control loops. In dynamic simulation, a delay module is used to characterize process delay, and a first-order module is used to increase the dead time of the process and equipment. The precise Tyreus-Luyben rule is used to sequentially perform relay feedback tests on the temperature and composition controllers for parameter tuning and refinement. The Tyreus-Luyben tuning formula is... In the formula, This represents the proportional gain of the oscillating loop. This represents the limiting gain of the oscillation curve. Indicates the integration time. This indicates the extreme period of the oscillation curve.

2. The multi-agent automatic control and simulation calculation method for a cluster distillation column according to claim 1, characterized in that... The separation performance of a single distillation unit within a cluster distillation column is simulated and calculated, specifically as follows: (1) Determine the physical property method and simulation unit module of the separation system: Use chemical process simulation software to determine the physical property method based on the physical characteristics of the reactive distillation process system; (2) Determine the optimal steady-state process parameters: Based on the system characteristics and actual process parameters of the reactive distillation process, a steady-state model of the distillation unit is established using chemical process simulation software. On the basis of the established steady-state model, the established steady-state model is simulated and tested. Based on the test results, the optimal structural parameters and operating parameters are determined. (3) Determine the control objectives and control degrees of freedom of the control system: Based on the steady-state design, and considering the temperature, gas-liquid phase flow rate and gas-liquid composition distribution of the distillation unit, determine the product quality and target product selectivity as the control objectives of the control system design according to the process requirements.

3. The multi-agent automatic control and simulation calculation method for a cluster distillation column according to claim 2, characterized in that... The chemical process simulation software used is Aspen. The specific unit modules selected in the simulation are: Heater module for heat exchangers, RadFrac module for towers, Mixer module for mixers, FSplit module for distributors, Flash2 module for flash tanks, and Pump module for booster pumps.

4. The multi-agent automatic control and simulation calculation method for a cluster distillation column according to claim 1, characterized in that... The multiple distillation units are coupled together to form a cluster distillation column, specifically: Ten distillation units are connected in parallel to form a cluster distillation column. The condenser and reboiler are set separately outside the unit. The material is uniformly fed into each distillation unit. The liquid material at the bottom of the cluster distillation column is vaporized in the reboiler and enters the distillation equipment through the vapor inlet to form rising vapor, which is evenly distributed into each distillation unit. The rising vapor leaves the cluster distillation column through the vapor outlet and enters the subsequent condenser to form condensate. A portion of the condensate is evenly distributed into each distillation unit of the distillation section through a reflux device. A portion of the material from the reboiler and condenser leaves the cluster distillation column at a concentration and flow rate that meets the separation process requirements, completing the entire distillation separation process.

5. The multi-agent automatic control and simulation calculation method for a cluster distillation column according to claim 1, characterized in that... The design of the multi-loop process control system for the cluster distillation column unit specifically includes: (1) The feed flow is controlled by flow control. The controller is a counteracting controller. The feed flow is controlled or adjusted by adjusting its set value. (2) The pressure at the top of the distillation column is controlled by adjusting the heat transfer of the condenser at the top of the column; (3) The composition-flow cascade control scheme of the cluster distillation column is used to adjust the reflux flow rate of the top product stream. A component controller is added to the top product stream of the cluster distillation column to control and maintain the concentration of the top product. The cascade control loop is formed by adjusting the set value of the flow controller. The component controller is the primary controller and the flow controller is the secondary controller. (4) The temperature of the sensitive trays in the distillation unit is controlled by adjusting the heat load of the reboiler in the column bottom; (5) The reboiler level in the cluster distillation column is controlled by adjusting the reboiler discharge flow rate.

6. The multi-agent automatic control and simulation calculation method for a cluster distillation column according to claim 1, characterized in that... The anti-interference performance of the cluster distillation column unit was tested, specifically as follows: For the design of a multi-loop automatic control system for a cluster distillation column, feed flow rate disturbance and feed composition disturbance are introduced into the dynamic simulation to analyze the dynamic response of the system. If the dynamic response meets the requirements of the control system, the design scheme is effective; otherwise, the controller parameters are readjusted and the loops with poor control performance in the control loop are modified until the dynamic response meets the system requirements.

Images