Multi-agent automatic control and analog calculation method for cluster type rectifying tower

Abstract

The invention discloses a multi-agent automatic control and analog calculation method for a cluster type rectifying tower, which comprises the following steps: columnar rectifying units in the cluster type rectifying tower are formed by separating a space area in the tower, the rectifying units are uniformly distributed and are mutually independent, and any rectifying units have no mass transfer; carrying out analog calculation on the separation performance of a single rectification unit; a plurality of rectification units are coupled to form a cluster type rectification tower, and the device is subjected to multi-loop process control system design; designing an intelligent agent unit controller according to the multi-coupling characteristics of the cluster type rectifying tower, and realizing uniform distribution and state consistency of each unit fluid in the cluster type rectifying tower; and carrying out dynamic simulation calculation on the multi-loop process control system, adjusting and optimizing parameters of the controller, and testing the anti-interference performance. Key operation parameters of a cluster type rectifying tower device are simulated and analyzed through a simulation calculation technology, the optimal rectifying condition is determined, a multi-agent automatic control system is designed, and a technical basis is provided for industrialization of the cluster type rectifying tower.

Description

Technical Field

[0001] The invention relates to the technical field of chemical technology application, and in particular to a multi-agent automatic control and simulation calculation method for a cluster distillation tower. Background Art

[0002] At present, distillation accounts for more than 70% of the chemical separation process. Industrial distillation tower equipment is large in size and large in number. The materials stay in the tower equipment for a long time and have a large storage capacity. Especially for difficult-to-separate systems with similar relative volatility and azeotropic systems, in order to achieve the separation effect, the number of theoretical plates often reaches hundreds, which requires the construction of distillation towers close to 100 meters in height, and the safety risks associated with it also increase dramatically. The development of high-efficiency fillers, new tower plates and internal components can theoretically improve the separation efficiency, but there are no significant improvement measures for distillation devices currently used in industrial applications.

[0003] Invention patent (a cluster distillation tower, CN202110144285.6). The centimeter-scale distillation units in the cluster distillation tower are filled with high-efficiency fillers, and the container-type distillation equipment is constructed by integrating the distillation cluster array, which greatly reduces the liquid holding volume in the distillation unit and greatly reduces the safety risk. The cluster distillation equipment can realize the industrial separation of difficult-to-separate system materials with a relative volatility close to 1, without the need for a distillation tower up to tens of meters, greatly reducing the equipment investment cost.

[0004] However, during the separation process, the cluster distillation tower device's operating efficiency, target quality of the treated mixture, economy and safety of production operations are all affected by many factors. Coupled with the complexity of the coupling between distillation units, this will lead to the cluster distillation tower device having great uncertainty and complexity. Summary of the invention

[0005] In view of this, the present invention provides a multi-agent automatic control and simulation calculation method for a cluster distillation tower, develops multi-agent automatic control technology, realizes integrated precision control of "stack / cluster" array integration, and improves equipment reliability. This method provides a technical basis for achieving sustainable and efficient cluster distillation tower production.

[0006] Technical Solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

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

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

[0010] 1) Determine the physical property method of the separation system and the unit module of the simulation: Use chemical process simulation software to determine the physical system method according to the physical system characteristics of the reaction distillation process system. The unit modules selected in the simulation include: Heater module for heat exchanger, RadFrac module for tower, Mixer module for mixer, FSplit module for splitter, Flash2 module for flash tank, and Pump module for booster pump.

[0011] (2) Determine the optimal steady-state process parameters: Based on the physical characteristics and actual process parameters of the reactive distillation process system, use chemical process simulation software to establish a steady-state model of the distillation unit. Based on 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, the product quality and target product selectivity are reasonably determined as the control objectives of the control system design according to the process requirements, with respect to the distillation unit temperature, gas-liquid phase flow rate, and gas-liquid composition distribution;

[0013] (2) Multiple distillation units are coupled and combined to form a cluster distillation tower, specifically:

[0014] Ten distillation units are connected in parallel to form a cluster distillation tower. The condenser and reboiler are only set up separately outside the device, and the materials enter each distillation unit evenly. After the material liquid at the bottom of the cluster distillation tower is vaporized by the reboiler, it enters the distillation equipment from the vapor phase inlet to form rising steam, and is evenly distributed into each distillation unit. The rising steam leaves the cluster distillation tower from the vapor phase outlet and enters the subsequent condenser to form condensate. The obtained part of the condensate is evenly distributed into each distillation unit of the distillation section through the reflux device. Part of the material in the reboiler and condenser leaves the cluster distillation tower device at an appropriate concentration and flow rate, completing the entire distillation separation process.

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

[0016] 1) The feed flow is flow controlled. The controller works in a reverse manner, and the feed flow is controlled or adjusted by adjusting its set value.

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

[0018] 3) The composition-flow cascade control scheme of the cluster distillation tower is used to adjust the tower top logistics reflux flow rate, and a component controller is added to the tower top product stream of the cluster distillation tower to control and maintain the tower top product concentration, and a 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.

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

[0020] 5) The liquid level of the reboiler in the cluster distillation tower device is controlled by adjusting the discharge flow rate of the tower bottom.

[0021] (4) Design an intelligent unit controller based on the multi-coupling characteristics of the cluster distillation tower to achieve consistency in fluid distribution within the unit, specifically:

[0022] Design a control input feed flow u based on the relative states between the agents. i (t), so that the MAN described by the linear second-order dynamics of a single agent is consistent. The definition of consistency is as follows:

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

[0024]

[0025] Then MAN achieves consistency. In order to achieve consistency, the consistency control input of MAN is designed as

[0026]

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

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

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

[0030]

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

[0032] (5) Optimize the controller parameters for the cluster distillation tower device, specifically:

[0033] The proportional-integral-derivative (PID) controller is used to design the control loop. Deadtime is introduced into the temperature and component control loops. In the dynamic simulation, the delay module is used to characterize the process delay, and the first-order module is used to increase the deadtime of the process and equipment. The temperature and component controllers are sequentially tested with relay feedback tests using the precise Tyreus-Luyben rule to perform parameter tuning and setting. The Tyreus-Luyben tuning formula is:

[0034]

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

[0036] For the multi-loop automatic control system designed for cluster distillation tower, feed flow disturbance and feed composition disturbance are introduced in the dynamic simulation respectively to analyze the dynamic response of the system; if the dynamic response meets the control system requirements, the design scheme is valid, otherwise, the controller parameters are readjusted and the loops with poor control effects in the control loop are modified until the dynamic response meets the system requirements.

[0037] Beneficial Effects

[0038] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0039] Using simulation technology, we simulate and analyze a series of parameters such as the feed state, material feed rate, distillation temperature, pressure, etc. of each distillation unit in the cluster distillation tower device to determine the optimal distillation conditions. We design a multi-loop process control system and perform dynamic simulation calculations on the cluster distillation tower to provide a technical basis for the industrial cluster distillation tower device and guide practice through theory. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of a process simulation of a cluster distillation tower of the present invention;

[0041] Figure 2 A schematic diagram of a multi-agent automatic control system for a cluster distillation tower of the present invention;

[0042] Figure 3 The dynamic response diagram of the system after adding feed flow rate and feed disturbance in implementation case 1 ((1) o-chlorotoluene content under feed flow disturbance; (2) o-chlorotoluene content under feed flow disturbance);

[0043] Figure 4 The dynamic response diagram of the system after adding feed flow rate and feed disturbance in implementation case 2 ((1) o-chlorobenzyl chloride content under feed flow disturbance; (2) o-chlorobenzyl chloride content under feed flow disturbance);

[0044] Figure 5 The dynamic response diagram of the system after adding feed flow rate and feed disturbance in implementation case 3 ((1) p-chlorobenzyl chloride content under feed flow disturbance; (2) p-chlorobenzyl chloride content under feed flow disturbance); DETAILED DESCRIPTION

[0045] Implementation Case 1: Simulation case of o- / p-chlorotoluene separation process in cluster distillation tower device

[0046] Step 1: Use simulation software to determine the physical property method of the separation system and the unit module of the simulation

[0047] Using chemical process simulation software (Aspen is used in this embodiment, and in other embodiments, the Mystic Simulation Software and APEX software can also be used to achieve simulation), the physical system method is determined to be the PR (Peng-Robinson) state equation according to the physical system characteristics of the reaction distillation process system. The unit modules selected in the simulation include: Heater module for heat exchanger, RadFrac module for tower, Mixer module for mixer, FSplit module for splitter, Flash2 module for flash tank, and Pump module for booster pump.

[0048] Step 2: Taking a single distillation unit as the object, simulate the distillation separation process:

[0049] According to the gas wave equilibrium constant, boiling point, relative volatility, and other system characteristics of the distillation process system and the actual process parameters, the chemical process simulation software is used to establish the steady-state model of the o- / p-chlorotoluene separation process. Figure 1 As shown in Table 1 (different steady-state mechanism models of distillation towers are built according to different separation systems), on the basis of the established steady-state model of the process flow, the optimal steady-state process structure parameters and operating parameters of the multi-unit reaction distillation device are determined (the built steady-state model is simulated and tested in the software, and the optimal structure parameters and operating parameters are determined according to the test results).

[0050] Step 3: Determine the control objectives and control degrees of freedom for the o- / p-chlorotoluene separation control system

[0051] On a steady-state basis, according to the process requirements, the product quality and target product selectivity are reasonably determined as the control objectives of the control system design, targeting the tower plate temperature, gas-liquid phase flow rate and gas-liquid composition distribution in the tower: According to the process steady-state simulation process, the control freedom of the control system (i.e., the controlled variable) is determined: In terms of production indicators, when designing the control system for the o-chlorotoluene / para-chlorotoluene separation process of the cluster distillation tower device, it is required that the o-chlorotoluene content (MoleFraction) is stable at around 99.5%. In the o-chlorotoluene / para-chlorotoluene separation process of the cluster distillation tower device, there are multiple controlled variables used to control the entire process. They are the pressure of each distillation unit, the distillation unit temperature, the reflux rate, the feed rate, the subcooling temperature, and the purity of the o-chlorotoluene / para-chlorotoluene produced.

[0052] Table 1

[0053]

[0054]

[0055] Step 4: coupling and combining multiple distillation units to form a cluster distillation tower;

[0056] Based on the simulation of a single distillation unit, 10 distillation units (RadFrac1-RadFrac6) are connected in parallel to build a cluster distillation tower simulation device. The material (ortho- / para-chlorotoluene mixture) enters each distillation unit evenly; the material liquid at the bottom of the cluster distillation tower is vaporized by the reboiler (Heater2), enters the distillation equipment from the vapor phase inlet to form rising steam, and is evenly distributed into each distillation unit. The rising steam leaves the cluster distillation tower from the vapor phase outlet and enters the subsequent condenser (Heater1) to form a condensate. The obtained partial condensate is evenly distributed into each distillation unit of the distillation section through the reflux device. Part of the material in the reboiler and condenser leaves the cluster distillation tower device at a concentration and flow rate that meets the separation requirements, completing the entire ortho- / para-chlorotoluene distillation and separation process.

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

[0058] Table 2

[0059] Cluster distillation tower operating parameters Numeric Simulated pressure (mmHg) 92 Tower top stream temperature (℃) 92.28 Top distillation volume (kg / h) 5.32 Molar reflux ratio 35.4 Tower bottom logistics temperature (℃) 115.06 Tower bottom distillation volume (kg / h) 5.32 Content of o-chlorotoluene in the top discharge 99.5 Content of p-chlorotoluene in the bottom discharge 99.5

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

[0061] Step 5: Design a multi-loop process control system for the cluster distillation tower device for separation of o-chlorotoluene / p-chlorotoluene, and select temperature, pressure, and liquid level process variables closely related to the production process for design of control loops, such as Figure 2 shown.

[0062] 1) The feed flow is flow controlled. The controller works in a reverse manner, and the feed flow is controlled or adjusted by adjusting its set value.

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

[0064] 3) The composition-flow cascade control scheme of the distillation tower is used to adjust the reflux flow of the tower top logistics, and a component controller is added to the tower top product stream of the distillation tower to control and maintain the concentration of o-chlorotoluene therein, and a 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.

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

[0066] 5) The liquid level of the tower bottom reboiler in the cluster device is controlled by regulating the tower bottom discharge flow rate.

[0067] Step 6: Design an intelligent unit controller based on the multi-coupling characteristics of the cluster distillation tower for o-chlorotoluene / p-chlorotoluene separation to achieve consistency in fluid distribution within the unit, specifically:

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

[0069]

[0070] Among them, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij represents the adjacency matrix between the i-th and j-th agents, K r(t) is the control gain to be designed, is the switching signal, which determines the mode of communication topology, r(t) is a hidden Markov chain that depends on the Markov process, and g(x i (t)) is the controller disturbance caused by the system switching topology, which is assumed to be bounded.

[0071] Step 7: Optimizing controller parameters for the cluster distillation tower device for separating ortho / para-chlorotoluene, specifically:

[0072] 1) The proportional gain Kc of all liquid level controllers is set to 2, and the integration time is 9999 minutes.

[0073] 2) The proportional gain Kc of all pressure controllers is set to 5 and the integration time is 20 minutes.

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

[0075] Deadtime is introduced into the temperature and composition control loops: 1 minute and 2 minutes in the distillation tower and reactor temperature control loops, and 3 minutes in the composition control loop. In the dynamic simulation, the delay module is used to characterize the process delay, and the first-order module is used to increase the deadtime of the process and equipment. The temperature and composition controllers are sequentially tested with relay feedback using the precise Tyreus-Luyben rule to perform parameter tuning and setting.

[0076] Taking the temperature control loop in the distillation unit RadFrac1 as an example, it is first set as a closed loop and the disturbance amplitude is set to 5%. Then the relay feedback test is carried out to obtain the dynamic response and oscillation curve of the control loop. The Tyreus-Luyben rule is then used to calculate the limit period Pu = 0.33856 and the limit gain K of the oscillation curve. n = 6 minutes. After the test, The proportional gain Kc of the corresponding loop is calculated to be 0.1058 and the integration time τ1 is 13.2 minutes.

[0077] Step 8: Disturbance operation and analysis of control system performance

[0078] Introduce disturbances, analyze the dynamic response performance indicators of the system, evaluate the effectiveness of the design scheme, and modify the control loop as needed to make the system meet the control requirements. Figure 3 The performance test shows that the control system has good control performance, can effectively overcome the disturbance 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 dichloride in a cluster distillation tower device

[0080] Step 1: Use simulation software to determine the physical property method of the separation system and the unit module of the simulation

[0081] The chemical process simulation software was used to determine the physical system method as the PR (Peng-Robinson) state equation according to the physical system characteristics of the reaction distillation process system. The unit modules selected in the simulation include: Heater module for heat exchanger, RadFrac module for tower, Mixer module for mixer, FSplit module for splitter, Flash2 module for flash tank, and Pump module for booster pump.

[0082] Step 2: Taking a single distillation unit as the object, simulate the distillation separation process:

[0083] According to the gas wave equilibrium constant, boiling point, relative volatility, and other system characteristics of the distillation process system and the actual process parameters, 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 reaction distillation device are determined.

[0084] Step 3: Determine the control target and control degree of freedom of the control system for the separation of o-chlorobenzyl chloride and o-chlorobenzylidene dichloride

[0085] On a steady-state basis, according to the process requirements, the product quality and target product selectivity are reasonably determined as the control objectives of the control system design, targeting the tower plate temperature, gas-liquid phase flow rate and gas-liquid composition distribution in the tower: According to the process steady-state simulation process, the control freedom of the control system (i.e., the controlled variable) is determined: In terms of production indicators, when designing the control system for the separation process of o-chlorobenzyl chloride and o-chlorobenzylidene dichloride in a cluster distillation tower device, it is required that the content of chlorobenzyl chloride (MoleFraction) is stable at around 95%. In the process flow of separation of o-chlorobenzyl chloride and o-chlorobenzylidene dichloride in a cluster distillation tower device, there are multiple controlled variables used to control the entire process flow. They are the pressure of each distillation unit, the temperature of the distillation unit, the reflux rate, the feed rate, the subcooling temperature, and the purity of the produced chlorobenzyl chloride.

[0086] Table 3

[0087]

[0088] Step 4: coupling and combining multiple distillation units to form a cluster distillation tower;

[0089] Based on the simulation of a single distillation unit, 10 distillation units (RadFrac1-RadFrac6) are connected in parallel to build a cluster distillation tower simulation device. The material (mixture of o-chlorobenzyl chloride and o-chlorobenzylidene dichloride) enters each distillation unit evenly; the material liquid at the bottom of the cluster distillation tower is vaporized by the reboiler (Heater2), enters the distillation equipment from the vapor phase inlet to form rising steam, and is evenly distributed into each distillation unit. The rising steam leaves the cluster distillation tower from the vapor phase outlet and enters the subsequent condenser (Heater1) to form a condensate. The obtained partial condensate is evenly distributed into each distillation unit of the distillation section through the reflux device. Part of the material from the reboiler and condenser leaves the cluster distillation tower device at a concentration and flow rate that meets the separation requirements, completing the entire o- / p-chlorotoluene distillation and separation process.

[0090] After the above simulation process is completed, the simulation results of important logistics are shown in Table 4:

[0091] Table 4

[0092] Cluster distillation tower operating parameters Numeric Simulated pressure (mmHg) 97 Tower top stream temperature (℃) 101.5 Top distillation volume (kg / h) 2.87 Molar reflux ratio 3.1 Tower bottom logistics temperature (℃) 113 Content of o-chlorobenzyl chloride in the top discharge 95

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

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

[0095] 1) The feed flow is flow controlled. The controller works in a reverse manner, and the feed flow is controlled or adjusted by adjusting its set value.

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

[0097] 3) The composition-flow cascade control scheme of the distillation tower is used to adjust the reflux flow of the tower top logistics, and a component controller is added to the tower top product stream of the distillation tower to control and maintain the concentration of o-chlorotoluene therein, and a 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.

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

[0099] 5) The liquid level of the tower bottom reboiler in the cluster device is controlled by regulating the tower bottom discharge flow rate.

[0100] Step 6: Design an intelligent unit controller based on the multi-coupling characteristics of the cluster distillation tower for separation of o-chlorobenzyl chloride and o-chlorobenzylidene dichloride to achieve consistency in fluid distribution within the unit, specifically:

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

[0102]

[0103] Among them, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij represents the adjacency matrix between the i-th and j-th agents, K r(t) is the control gain to be designed, is the switching signal, which determines the mode of communication topology, r(t) is a hidden Markov chain that depends on the Markov process, and g(x i (t)) is the controller disturbance caused by the system switching topology, which is assumed to be bounded.

[0104] Step 7: Optimizing controller parameters of a cluster distillation tower device for separating o-chlorobenzyl chloride and o-chlorobenzylidene dichloride, specifically:

[0105] 1) The proportional gain Kc of all liquid level controllers is set to 2, and the integration time is 9999 minutes.

[0106] 2) The proportional gain Kc of all pressure controllers is set to 5 and the integration time is 20 minutes.

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

[0108] Deadtime is introduced into the temperature and composition control loops: 1 minute and 2 minutes in the distillation tower and reactor temperature control loops, and 3 minutes in the composition control loop. In the dynamic simulation, the delay module is used to characterize the process delay, and the first-order module is used to increase the deadtime of the process and equipment. The temperature and composition controllers are sequentially tested with relay feedback using the precise Tyreus-Luyben rule to perform parameter tuning and setting.

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

[0110] Step 8: Disturbance operation and analysis of control system performance

[0111] Introduce disturbances, analyze the dynamic response performance indicators of the system, evaluate the effectiveness of the design scheme, and modify the control loop as needed to make the system meet the control requirements. Figure 4 The performance test shows that the control system has good control performance, can effectively overcome the disturbance in the production process, and meets the requirements of industrial production control.

[0112] Implementation Case 3: Simulation case of p-chlorobenzyl chloride / p-chlorobenzylidene dichloride separation process in cluster distillation tower device

[0113] Step 1: Use simulation software to determine the physical property method of the separation system and the unit module of the simulation

[0114] The chemical process simulation software was used to determine the physical system method as the PR (Peng-Robinson) state equation according to the physical system characteristics of the reaction distillation process system. The unit modules selected in the simulation include: Heater module for heat exchanger, RadFrac module for tower, Mixer module for mixer, FSplit module for splitter, Flash2 module for flash tank, and Pump module for booster pump.

[0115] Step 2: Taking a single distillation unit as the object, simulate the distillation separation process:

[0116] According to the gas wave equilibrium constant, boiling point, relative volatility, and other system characteristics of the distillation process system and the actual process parameters, a steady-state model of the p-chlorobenzyl chloride / p-chlorobenzylidene dichloride separation process 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 reaction distillation device are determined.

[0117] Step 3: Determine the control objectives and control freedom of the p-chlorobenzyl chloride / p-chlorobenzylidene dichloride separation control system

[0118] On a steady-state basis, according to the process requirements, the product quality and target product selectivity are reasonably determined as the control objectives of the control system design, targeting the tower plate temperature, gas-liquid phase flow rate and gas-liquid composition distribution in the tower: According to the process steady-state simulation process, the control freedom of the control system (i.e., the controlled variable) is determined: In terms of production indicators, when designing the control system for the p-chlorobenzyl chloride / p-chlorobenzylidene dichloride separation process of the cluster distillation tower device, it is required that the p-chlorobenzyl chloride content (MoleFraction) is stable at around 95%. In the p-chlorobenzyl chloride / p-chlorobenzylidene dichloride separation process of the cluster distillation tower device, there are multiple controlled variables used to control the entire process. They are the pressure of each distillation unit, the distillation unit temperature, the reflux rate, the feed rate, the subcooling temperature, and the purity of the produced p-chlorobenzyl chloride.

[0119] Table 5

[0120]

[0121] Step 4: coupling and combining multiple distillation units to form a cluster distillation tower;

[0122] Based on the simulation of a single distillation unit, 10 distillation units (RadFrac1-RadFrac10) are connected in parallel to build a cluster distillation tower simulation device. The material (p-chlorobenzyl chloride / p-chlorobenzylidene dichloride mixture) enters each distillation unit evenly; the material liquid at the bottom of the cluster distillation tower is vaporized by the reboiler (Heater2), enters the distillation equipment from the vapor phase inlet to form rising steam, and is evenly distributed into each distillation unit. The rising steam leaves the cluster distillation tower from the vapor phase outlet and enters the subsequent condenser (Heater1) to form a condensate. The obtained partial condensate is evenly distributed into each distillation unit of the distillation section through the reflux device. Part of the material from the reboiler and condenser leaves the cluster distillation tower device at a concentration and flow rate that meets the separation requirements, completing the entire ortho / p-chlorotoluene distillation separation process.

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

[0124] Table 6

[0125] Cluster distillation tower operating parameters Numeric Simulated pressure (mmHg) 97 Tower top stream temperature (℃) 110 Top distillation volume (kg / h) 3.62 Molar reflux ratio 1.9 Tower bottom logistics temperature (℃) 126.06 Content of p-chlorobenzyl chloride in the top discharge 95

[0126] It can be seen from the above table that the content of p-chlorobenzyl chloride in the top discharge is 95%, which satisfies the requirements of the p-chlorobenzyl chloride / p-chlorobenzylidene dichloride separation process.

[0127] Step 5: Design a multi-loop process control system for the cluster distillation tower device for separating p-chlorobenzyl chloride / p-chlorobenzylidene dichloride, and select the temperature, pressure, and liquid level process variables closely related to the production process for the design of the control loop, such as Figure 2 shown.

[0128] 1) The feed flow is flow controlled. The controller works in a reverse manner, and the feed flow is controlled or adjusted by adjusting its set value.

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

[0130] 3) The composition-flow cascade control scheme of the distillation tower is used to adjust the reflux flow of the tower top logistics, and a component controller is added to the tower top product stream of the distillation tower to control and maintain the concentration of o-chlorotoluene therein, and a 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.

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

[0132] 5) The liquid level of the tower bottom reboiler in the cluster device is controlled by regulating the tower bottom discharge flow rate.

[0133] Step 6: Design an intelligent unit controller based on the multi-coupling characteristics of the cluster distillation tower for the separation of p-chlorobenzyl chloride / p-chlorobenzylidene dichloride to achieve the consistency of fluid distribution inside the unit, specifically:

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

[0135]

[0136] Among them, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij represents the adjacency matrix between the i-th and j-th agents, K r(t) is the control gain to be designed, is the switching signal, which determines the mode of communication topology, r(t) is a hidden Markov chain that depends on the Markov process, and g(x i (t)) is the controller disturbance caused by the system switching topology, which is assumed to be bounded.

[0137] Step 7: Optimizing controller parameters of the cluster distillation tower device for separating p-chlorobenzyl chloride / p-chlorobenzylidene dichloride, specifically:

[0138] 1) The proportional gain Kc of all liquid level controllers is set to 2, and the integration time is 9999 minutes.

[0139] 2) The proportional gain Kc of all pressure controllers is set to 5 and the integration time is 20 minutes.

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

[0141] Deadtime is introduced into the temperature and composition control loops: 1 minute and 2 minutes in the distillation tower and reactor temperature control loops, and 3 minutes in the composition control loop. In the dynamic simulation, the delay module is used to characterize the process delay, and the first-order module is used to increase the deadtime of the process and equipment. The temperature and composition controllers are sequentially tested with relay feedback using the precise Tyreus-Luyben rule to perform parameter tuning and setting.

[0142] Taking the temperature control loop in the distillation unit RadFrac1 as an example, it is first set as a closed loop and the disturbance amplitude is set to 5%. Then the relay feedback test is carried out to obtain the dynamic response and oscillation curve of the control loop. The Tyreus-Luyben rule is then used to calculate the limit period Pu = 0.4786 and the limit gain K of the oscillation curve. u = 8 minutes. After the test, P U The proportional gain Kc of the corresponding loop is calculated to be 0.2563 and the integration time τ1 is 7.5 minutes.

[0143] Step 8: Disturbance operation and analysis of control system performance

[0144] Introduce disturbances, analyze the dynamic response performance indicators of the system, evaluate the effectiveness of the design scheme, and modify the control loop as needed to make the system meet the control requirements. Figure 5 The performance test shows that the control system has good control performance, can effectively overcome the disturbance 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 tower, wherein the columnar distillation units in the cluster distillation tower are separated by the space area in the tower, the distillation units are evenly distributed, and the distillation units are independent of each other, and there is 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 tower. A multi-loop process control system is designed for the cluster distillation tower device, and temperature, pressure, and liquid level process variables are selected as controlled variables. An intelligent unit controller is designed based on the multi-coupling characteristics of the cluster distillation tower to achieve uniform distribution of fluids in each unit of the cluster distillation tower and state consistency. Dynamic simulation calculations of the multi-loop process control system are performed to optimize the controller parameters and test the anti-interference performance.

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

3. The multi-agent automatic control and simulation calculation method of a cluster distillation tower according to claim 2 is characterized in that Aspen is selected as the chemical process simulation software. The unit modules selected in the simulation are: Heater module for heat exchanger, RadFrac module for tower, Mixer module for mixer, FSplit module for diverter, Flash2 module for flash tank and Pump module for booster pump.

4. The multi-agent automatic control and simulation calculation method of a cluster distillation tower according to claim 1 is characterized in that The plurality of distillation units are coupled and combined to form a cluster distillation tower, specifically: Ten distillation units are connected in parallel to form a cluster distillation tower, and the condenser and reboiler are only separately set outside the device, and the material enters each distillation unit evenly; the material liquid at the bottom of the cluster distillation tower is vaporized by the reboiler, enters the distillation equipment from the vapor phase inlet to form rising steam, and is evenly distributed into each distillation unit; the rising steam leaves the cluster distillation tower from the vapor phase outlet, enters the subsequent condenser to form condensate, and part of the condensate obtained is evenly distributed into each distillation unit of the distillation section through a reflux device; part of the material in the reboiler and the condenser leaves the cluster distillation tower device at a concentration and flow rate that meets the requirements of the separation process, completing the entire distillation separation process.

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

6. The multi-agent automatic control and simulation calculation method of a cluster distillation tower according to claim 5, characterized in that According to the multi-coupling characteristics of the cluster distillation tower, an intelligent unit controller is designed to achieve the consistency of fluid distribution inside the unit, specifically: The design goal is to design a control input u based on the relative states between agents. i (t), so that the MAN described by the linear second-order dynamics of a single agent is consistent. Therefore, the consistency control flow input is designed as Among them, x j (t) represents the neighbor status information, x i (t) represents the state information of the i-th agent, θ ij represents the adjacency matrix between the i-th and j-th agents, K r(t) is the control gain to be designed, is the switching signal, which determines the mode of communication topology, r(t) is a hidden Markov chain that depends on the Markov process, and g(x i (t)) is the controller disturbance caused by the system switching topology, which is assumed to be bounded.

7. The multi-agent automatic control and simulation calculation method of a cluster distillation tower according to claim 1, characterized in that The controller parameters of the cluster distillation tower device are optimized as follows: The proportional-integral-differential PID controller is used to design the control loop, and the dead time is introduced into the temperature and component control loops. In the dynamic simulation, the delay module is used to characterize the process delay, and the first-order module is used to increase the dead time of the process and equipment. The temperature and component controllers are sequentially subjected to the relay feedback test using the precise Tyreus-Luyben rule to perform parameter tuning and setting. The Tyreus-Luyben tuning formula is: τ1=2.2P U In the formula, K C represents the proportional gain of the shock loop, K U represents the limit gain of the oscillation curve, τ1 represents the integration time, P U Represents the limit period of the oscillation curve.

8. The multi-agent automatic control and simulation calculation method of a cluster distillation tower according to claim 1, characterized in that The test anti-interference performance of the cluster distillation tower device is as follows; For the multi-loop automatic control system designed for cluster distillation tower, feed flow disturbance and feed composition disturbance are introduced in the dynamic simulation respectively to analyze the dynamic response of the system; if the dynamic response meets the control system requirements, the design scheme is valid, otherwise, the controller parameters are readjusted and the loops with poor control effects in the control loop are modified until the dynamic response meets the system requirements.

Images