A carbon dioxide capture device

Abstract

The application provides a carbon dioxide capturing device, relates to the field of carbon dioxide adsorption, and comprises a total trachea, a distributor, an absorption tower group composed of multiple structural identical sub-towers, a flow control valve group, a sensor group, a regeneration tower, a lean-liquid rich-liquid heat exchanger and a controller. The controller dynamically adjusts the opening degrees of an air inlet adjusting valve and proportional adjusting valves of each sub-tower based on the distributor pressure, the air inlet flow of each sub-tower, the pressure in the tower and the carbon dioxide concentration at the outlet collected by the sensor, and dynamically adjusts the number of operating sub-towers and the flow distribution according to the upstream load change signal and in combination with the constraint of the high-efficiency area of the gas phase flow window of the sub-tower. The controller also integrates a prediction model based on a long short-term memory neural network, adopts a rolling time domain optimization algorithm, and minimizes the regeneration energy consumption under the condition of ensuring safe production. The application realizes the efficient, stable and energy-saving operation of the decarbonization process through the collaborative control of the high-efficiency group, the buffer group and the standby group.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of carbon dioxide adsorption, in particular to a carbon dioxide capture device. BACKGROUND

[0002] In the industrial processes of petroleum chemical industry, coal chemical industry and hydrogen production, etc., the shift gas rich in hydrogen and carbon dioxide is often produced through processes such as heavy oil gasification and water gas shift;

[0003] At present, the chemical absorption method (such as amine method) is widely used in industry for carbon dioxide capture, and the core equipment is a circulating system composed of an absorption tower and a regeneration tower; when the absorption tower and the absorption medium are determined, there is usually an optimal gas flow range, and the high-efficiency region of the gas phase flow window. Within this window, the gas-liquid two-phase flow in the tower is ideal, the contact is sufficient, which can ensure that the carbon dioxide capture efficiency is maintained at a high level, and the pressure drop in the tower is controlled within the safe range allowed by the equipment. The traditional absorption tower usually adopts a single tower or a fixed parallel tower structure, which has the following disadvantages:

[0004] When a single large absorption tower is used, the design is usually optimized for the rated load. When the gas load is lower than the lower limit of the high-efficiency region, the gas velocity in the tower is too low, the mass transfer efficiency is significantly reduced, and the unit processing energy consumption of the regeneration system is sharply increased, causing energy waste. When the load exceeds the upper limit of the high-efficiency region, the pressure drop in the tower increases sharply, the gas-liquid contact time is insufficient, which easily leads to the carbon dioxide concentration exceeding the standard at the outlet, and even causes flooding. The multi-tower parallel scheme generally adopts a start-stop control strategy: when the load increases, the standby tower is started directly; when the load decreases, the running tower is isolated. This strategy cannot realize fine distribution of the load of each tower. Under non-rated total load, although the system can operate, the gas flow of each tower is often outside its high-efficiency region, either running at low load and low efficiency, or running at high load and limit, which is not easy to maintain the overall energy efficiency of the equipment at a relatively optimal level, and the frequent start-stop of the tower also increases the equipment wear and tear. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a carbon dioxide capture device, which realizes efficient, stable and energy-saving operation of the decarbonization process through the coordinated control of the high-efficiency group, the buffer group and the standby group, and is suitable for industrial scenes such as hydrogen production with frequent load fluctuations.

[0006] To solve the above technical problems, the technical scheme of the present application is as follows:

[0007] A carbon dioxide capture device, comprising a total gas pipe, a regeneration tower and a lean-rich liquid heat exchanger, the total gas pipe being provided with an inlet gas regulating valve, comprising:

[0008] An absorption tower group composed of a plurality of structurally identical sub-towers;

[0009] a distributor connected to the main pipe through a pipe for receiving the shifted gas from upstream, and the sub-towers are connected to the distributor through branch pipes;

[0010] a flow control valve group comprising proportional regulating valves arranged on each of the branch pipes for independently controlling the flow of shifted gas into the corresponding sub-tower;

[0011] a sensor group for collecting in real time the data of the gas inlet flow, internal pressure and carbon dioxide concentration in the gas outlet of each of the sub-towers, and the pressure data inside the distributor;

[0012] a controller communicatively connected to the sensor group, the flow control valve group and the gas inlet regulating valve, and configured to:

[0013] based on the pressure data inside the distributor collected by the sensor group, the opening degree of the gas inlet regulating valve is adjusted to maintain the pressure inside the distributor at a preset pressure value;

[0014] based on the gas inlet flow data of each of the sub-towers collected by the sensor group, the opening degree of the proportional regulating valve corresponding to the sub-tower is independently adjusted to control the flow of shifted gas into each of the running sub-towers at its preset flow set value;

[0015] based on the upstream load change signal and the tower internal pressure data and the corresponding carbon dioxide concentration data in the gas outlet of the sub-tower collected by the sensor group, the number of sub-towers in the running state in the absorption tower group and the corresponding gas inlet flow set value are dynamically adjusted to make the running sub-towers work in the high efficiency area of the corresponding gas phase flow window, wherein the high efficiency area refers to the corresponding gas inlet flow range of the sub-tower under the condition that the carbon dioxide capture efficiency is not less than 95% and the tower internal pressure drop is not more than 30 kPa.

[0016] Further, the controller divides the absorption tower group into high efficiency group, buffer group and standby group according to the control logic;

[0017] The high efficiency group is composed of the minimum number of sub-towers required to run to handle the rated gas volume, and is used to continuously run at high efficiency point under rated operating conditions;

[0018] The buffer group is composed of at least one sub-tower, and is used to limit the load change when the load fluctuates to maintain the stable operation of the high efficiency group under rated operating conditions;

[0019] The standby group is composed of at least one sub-tower, and is in cold standby or hot standby state;

[0020] The controller pre-stores a first reference flow value and a second reference flow value, the first reference flow value corresponding to the upper limit of the high-efficiency region of the gas flow window, and the second reference flow value corresponding to the lower limit of the high-efficiency region of the gas flow window;

[0021] The controller is configured to control the high-efficiency group to operate at the first reference flow value as the flow set value, control the buffer group to operate at the second reference flow value as the flow set value, and maintain the standby group in the standby state when the total gas amount processed by the device corresponds to the rated load.

[0022] Further, the controller pre-stores a first opening threshold and a second opening threshold, and is configured to perform the following dynamic adjustment strategy:

[0023] When receiving an upstream load change signal indicating an increase in load, the flow set value of the sub-tower of the buffer group is preferentially increased;

[0024] When the opening of the proportional regulating valve of all sub-towers of the buffer group reaches or exceeds the first opening threshold, it is determined that the buffer group has insufficient adjustment margin, and at this time, the flow set value of the sub-tower of the high-efficiency group is increased;

[0025] If the load continues to increase, the flow set values of the high-efficiency group and the buffer group both reach the first reference flow value, and the opening of the proportional regulating valve of any sub-tower in the buffer group and the high-efficiency group is above the pre-set first opening threshold, the controller activates one or more sub-towers from the standby group and divides the corresponding sub-towers into the buffer group;

[0026] When receiving an upstream load change signal indicating a decrease in load, the flow set value of the sub-tower of the buffer group is preferentially decreased;

[0027] When the opening of the proportional regulating valve of the sub-tower of the buffer group is reduced to below the pre-set second opening threshold, it is determined that the adjustment capacity of the buffer group is close to the lower limit, and the controller isolates the sub-towers in the buffer group in sequence and divides the corresponding isolated sub-towers into the standby group.

[0028] Further, the controller is configured to:

[0029] Based on the real-time data collected by the sensor group, the instantaneous operating state parameters of each sub-tower are calculated, including the opening of the proportional regulating valve;

[0030] Based on the real-time data, the controller predicts the performance parameter trend of each sub-tower in the future period of time through the built-in prediction model based on the long short-term memory neural network;

[0031] The instantaneous operating state parameters and the predicted trend are combined to make a joint decision:

[0032] When the proportional regulating valve opening degree does not reach the first opening threshold value, and the predicted trend indicates that the outlet carbon dioxide concentration will exceed the preset concentration alarm value at any time within the predicted time window, a preventive regulation is started to adjust the load distribution in advance;

[0033] When the proportional regulating valve opening degree has reached or exceeded the first opening threshold value, the regulation is performed through the dynamic adjustment strategy.

[0034] Further, the first opening threshold value is set in the range of 80% to 95%, and the second opening threshold value is set in the range of 15% to 30%.

[0035] Further, the performance parameter trend predicted by the prediction model includes the pressure drop change rate of the sub-tower or the outlet carbon dioxide concentration change trend, and the predicted time window is 5 to 15 minutes.

[0036] Further, the prediction model is a data-driven model constructed based on a long short-term memory neural network, the input of which includes time series data of the inlet flow, the tower pressure and the outlet carbon dioxide concentration of each sub-tower in the past 30 minutes, and the output is the predicted value of the outlet carbon dioxide concentration of each sub-tower in the future 5 to 15 minutes.

[0037] Further, the preventive regulation adopts a receding horizon optimization algorithm, the optimization algorithm is based on a preset mapping relationship between the regeneration energy consumption and the sub-tower load, and under the constraint that the outlet carbon dioxide concentration does not exceed the preset limit value, the optimal inlet flow distribution scheme of each sub-tower is calculated, and the mapping relationship is a function relationship of the regeneration energy consumption with respect to the sub-tower inlet flow and the corresponding rich liquid flow and carbon dioxide concentration obtained by fitting historical operation data.

[0038] Further, the controller is configured to adjust the weights of the optimization target and the constraint condition according to the relationship between the predicted outlet carbon dioxide concentration and the preset concentration alarm value when the preventive regulation is performed:

[0039] When the predicted outlet carbon dioxide concentration is lower than the preset concentration alarm value within the future predicted time window, the energy consumption optimization is prioritized;

[0040] When the predicted outlet carbon dioxide concentration will exceed the preset concentration alarm value at any time within the future predicted time window, the concentration constraint is prioritized, and the energy consumption optimization is temporarily suspended.

[0041] The above scheme of the present application at least has the following beneficial effects:

[0042] The above scheme of the present application realizes real-time monitoring of the pressure in the sub-tower and the carbon dioxide concentration at the gas outlet by the sensor group, dynamically adjusts the gas inlet flow and the number of operating sub-towers, ensures that the whole sub-tower in operation is in the efficient area of the gas flow window, and maintains a certain adjustment margin, so as to maintain the high-efficiency and stable operation of the device;

[0043] By dividing the absorption tower group into high-efficiency groups, buffer groups and standby groups according to the control logic, reserving the adjustment capacity under the rated working condition, preferentially adjusting the buffer group when the load changes, and maximizing the operation of the high-efficiency group at the optimal working point, the overall operation stability of the system is further improved;

[0044] Based on the multivariate time series data of the past 30 minutes, the long short-term memory neural network is used to accurately predict the carbon dioxide concentration change trend of each sub-tower outlet in the next 5 to 15 minutes; when the prediction exists the risk of exceeding the standard and the adjustment valve still has a margin, the preventive load redistribution is started in advance to reduce the risk of concentration overshoot;

[0045] By fitting the function model of the regeneration energy consumption with respect to the gas inlet flow, the rich liquid flow and the carbon dioxide concentration based on the historical data, the minimum regeneration tower steam consumption is taken as the target in the rolling optimization, while the outlet carbon dioxide concentration is strictly constrained not to exceed the standard; and the priority of safety priority or energy efficiency priority is dynamically adjusted according to the predicted risk level, so as to save energy while ensuring safety in production. BRIEF DESCRIPTION OF DRAWINGS

[0046] Fig. 1 is the overall structure schematic diagram provided by the present application.

[0047] Fig. 2 is the flow chart of the controller control logic in the present application.

[0048] In the figure: 100, total gas pipe; 200, gas inlet adjusting valve; 300, distributor; 400, sub-tower; 500, proportional adjusting valve; 600, one-way valve body; 700, spraying mechanism; 800, branch pipe; 900, gas outlet pipe. DETAILED DESCRIPTION

[0049] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0050] As Figs. 1-2 shown, the embodiment of the present application proposes a carbon dioxide capture device, which comprises a total gas pipe 100, a regeneration tower and a rich-liquid and lean-liquid heat exchanger, and the total gas pipe 100 is provided with a gas inlet adjusting valve 200, comprising:

[0051] a plurality of sub-towers 400;

[0052] a distributor 300 connected to the main gas pipe 100 through a pipe for receiving the shift gas from upstream, and the sub-towers 400 are connected to the distributor 300 through branch pipes 800;

[0053] a flow control valve group including proportional regulating valves 500 arranged on each branch pipe 800 for independently controlling the flow of shift gas into the corresponding sub-tower 400;

[0054] a sensor group for collecting real-time data of the gas flow into each sub-tower 400, the internal pressure, the carbon dioxide concentration at the gas outlet, and the pressure inside the distributor 300;

[0055] a controller in communication connection with the sensor group, the flow control valve group, and the gas inlet regulating valve 200, and configured to:

[0056] based on the pressure data inside the distributor 300 collected by the sensor group, the opening degree of the gas inlet regulating valve 200 is adjusted to maintain the pressure inside the distributor 300 at a preset pressure value;

[0057] based on the gas flow data of each sub-tower 400 collected by the sensor group, the opening degree of the proportional regulating valve 500 of the corresponding sub-tower 400 is independently adjusted to control the flow of shift gas into each running sub-tower 400 at its preset flow set value;

[0058] based on the upstream load change signal and the tower internal pressure data and the corresponding carbon dioxide concentration data at the gas outlet of the sub-tower 400 collected by the sensor group, the number of sub-towers 400 in the running state in the absorption tower group and the corresponding gas flow set value are dynamically adjusted to make the running sub-tower 400 work in the efficient area of the corresponding gas phase flow window, wherein the efficient area refers to the corresponding gas flow range of the sub-tower 400 under the condition that the carbon dioxide capture efficiency is not less than 95% and the tower internal pressure drop is not more than 30 kPa.

[0059] In the embodiment of the present application, in the hydrogen production process stage of petroleum chemical industry, the production raw material is heavy oil, and the shift gas rich in hydrogen and carbon dioxide is generated through gasification and shift reaction, which needs to be treated by the decarburization device of the present application before entering the PSA hydrogen extraction unit;

[0060] the main gas pipe 100 is the main pipe of the shift gas from the upstream shift section, and the gas inlet regulating valve 200 is arranged thereon for adjusting the total shift gas amount entering the entire capture device for treatment;

[0061] The absorption tower group is composed of a plurality of sub-towers 400 which are completely same in structure and can be independently operated in parallel, each of the sub-towers 400 is internally provided with a spraying mechanism 700 and a packing mechanism, the bottom of the sub-tower 400 is provided with a rich liquid outlet, the top of the sub-tower 400 is provided with a gas outlet, the gas outlet is connected with the inside of a gas outlet pipe 900 through a one-way valve body 600, one end of the gas outlet pipe 900 is closed, the middle of the sub-tower 400 is provided with a lean liquid inlet, which is prior art and will not be described in detail, the design parameters of the tower diameter, packing height and liquid-gas ratio of each sub-tower 400 are the same, so as to ensure that each sub-tower 400 can have the same gas flow window high efficiency area; in the embodiment, the gas flow window high efficiency area is that when the conversion gas inlet flow of the sub-tower 400 is in a certain range, the gas-liquid two-phase in the tower is fully contacted, the carbon dioxide capture efficiency can be stably maintained above 95%, and the pressure drop in the tower does not exceed the designed allowable value (such as 30 kPa), in the embodiment, the upper limit and the lower limit of the gas flow window high efficiency area can be determined through the process design calculation of the sub-tower 400 combined with the actual calibration test, and stored as the preset parameters of the controller;

[0062] The distributor 300 is a tubular structure with both ends closed, and a baffle plate is arranged inside the distributor 300, the surface of the baffle plate is uniformly provided with gas holes, the total gas pipe 100 and the branch pipe 800 are respectively located on the two sides of the baffle plate, so as to make the conversion gas entering the inside of the distributor 300 uniformly distributed after passing through the baffle plate, the distributor 300 is connected with the inside of the total gas pipe 100 through a pipeline, and is used for receiving and distributing the conversion gas to the corresponding sub-tower 400, and each sub-tower 400 is communicated with the distributor 300 through an independent branch pipe 800.

[0063] The flow control valve group is provided with an electric proportional regulating valve 500 on each branch pipe 800, the opening degree of the electric proportional regulating valve 500 can be independently adjusted by the controller, so as to accurately control the conversion gas flow entering the corresponding sub-tower 400;

[0064] The sensor group includes the following sensors:

[0065] The pressure sensor installed on the distributor 300 is used for monitoring the pressure inside the distributor 300;

[0066] The flow sensor installed on the branch pipe 800 of each sub-tower 400 is used for measuring the inlet gas flow;

[0067] The pressure sensor installed in each sub-tower 400 is used for monitoring the operating pressure in the tower;

[0068] The online carbon dioxide concentration analyzer installed on the purified gas outlet of each sub-tower 400 is used for monitoring the carbon dioxide concentration in the outlet gas in real time.

[0069] The controller, which is an industrial programmable logic controller (PLC) or distributed control system (DCS), is connected to all the sensors, proportional regulating valves 500 and inlet regulating valves 200 on the total gas pipe 100. The controller pre-stores the high-efficiency flow range, control logic and operation program of each sub-tower 400.

[0070] The lean-rich liquid system connects the lean liquid inlet of each sub-tower 400 to the lean liquid main pipe through a branch pipe 800 with a shut-off valve, and connects the rich liquid outlet to the rich liquid main pipe through a branch pipe 800 with a shut-off valve. When one sub-tower 400 is activated, the corresponding lean liquid and rich liquid shut-off valves are opened synchronously. When it is isolated, the shut-off valves are closed to prevent liquid from flowing or the tower from running dry. The rich liquid is preheated in the lean-rich liquid heat exchanger and then sent to the regeneration tower. The regenerated lean liquid is cooled in the heat exchanger and then sent back to the lean liquid main pipe by the lean liquid pump for recycling by each operating sub-tower 400.

[0071] The controller reads the data of the pressure sensor of the distributor 300 in real time and compares it with the pre-set pressure value (e.g. 0.35 MPa). If the measured pressure deviates from the set value, the controller adjusts the opening of the inlet regulating valve 200 on the total gas pipe 100 to maintain stable pressure of the distributor 300 and ensure consistent pressure at the inlet of each branch pipe 800.

[0072] For each sub-tower 400 in operation, the controller sets a flow set value for it (which is within its high-efficiency range). The controller adjusts the opening of the proportional regulating valve 500 on the corresponding branch pipe 800 based on the feedback from the flow sensor on the branch pipe 800 to stabilize the actual inlet flow at the set value.

[0073] The controller receives upstream load change signals, which can be a total gas amount set value change instruction issued by the upstream shift section or a load change trend automatically identified by the controller based on the historical trend of total inlet gas amount.

[0074] During the operation of the device, the equipment operating environment is basically constant, and the processing capacity of the sub-tower 400 is mainly affected by the gas phase load. The controller continuously monitors the internal pressure of each sub-tower 400 and the carbon dioxide concentration of the outlet gas.

[0075] When the inlet flow is too large, the gas velocity in the tower increases, resulting in:

[0076] Shortened gas-liquid contact time, insufficient carbon dioxide removal, increased carbon dioxide concentration at the outlet of the sub-tower 400, increased pressure drop in the tower, which may approach the safety limit of the equipment. Therefore, the internal pressure of the tower and the carbon dioxide concentration at the outlet of the sub-tower 400 can be used as reliable indicators to determine whether the sub-tower 400 is approaching the upper limit of the high-efficiency range.

[0077] When the upstream load increases, the controller first gradually increases the flow set value of the current operating sub-tower 400 without adding a sub-tower 400; if the pressure in any sub-tower 400 exceeds the preset threshold (such as 28 kPa), or the carbon dioxide concentration at the gas outlet of the sub-tower 400 exceeds the preset warning value (such as 2.0 vol%), it is determined that the tower has reached the upper limit of the high-efficiency zone;

[0078] At this time, the controller starts a standby sub-tower 400: first open its lean liquid and rich liquid shut-off valves to establish liquid circulation; then slowly open its gas proportional regulating valve 500, and the initial flow set value is set to the lower limit of the high-efficiency zone (such as 60% of the design rated flow); the controller recalculates the total load distribution and adjusts the flow set values of all operating sub-towers 400 to within their respective high-efficiency zones to ensure overall efficient operation.

[0079] When the upstream load decreases, the controller preferentially reduces the flow set values of the operating sub-towers 400; if the total load decreases to the point that the flow of most sub-towers 400 is below the lower limit of the high-efficiency zone (such as 40% of the rated flow), and the outlet carbon dioxide concentration is significantly lower than the control standard (such as less than 0.5 vol%), indicating that the absorption capacity is excessive; the controller selects a sub-tower 400 for isolation: first close its gas proportional regulating valve 500, and then close its lean liquid and rich liquid shut-off valves after the gas in the tower is exhausted; the flow set values of the remaining sub-towers 400 are correspondingly increased to return their operating points to the high-efficiency zone to avoid energy efficiency degradation at low load.

[0080] The controller divides the absorption tower group into a high-efficiency group, a buffer group, and a standby group according to the control logic;

[0081] The high-efficiency group is composed of the minimum number of sub-towers 400 required to handle the rated gas volume and is used to continuously operate at the high-efficiency point under rated operating conditions;

[0082] The buffer group is composed of at least one sub-tower 400 and is used to accommodate load changes to a limited extent to maintain stable operation of the high-efficiency group under rated operating conditions;

[0083] The standby group is composed of at least one sub-tower 400 and is in a cold standby or hot standby state;

[0084] The controller pre-stores a first reference flow value and a second reference flow value, the first reference flow value corresponding to the upper limit of the high-efficiency zone of the gas phase flow window, and the second reference flow value corresponding to the lower limit of the high-efficiency zone of the gas phase flow window;

[0085] The controller is configured to control the high-efficiency group to operate at the first reference flow value as the flow set value when the total gas volume handled by the device corresponds to the rated load, while controlling the buffer group to operate at the second reference flow value as the flow set value, and maintaining the standby group in standby state.

[0086] In the embodiment of the present application, the absorption tower group is composed of four sub-towers 400, namely TA1, TA2, TA3 and TA4. The controller divides the absorption tower group into three functional groups according to the preset logic: the high-efficiency group, the buffer group and the standby group.

[0087] The high-efficiency group is composed of the sub-towers 400 that bear the basic load under the rated operating condition, so as to ensure the carbon dioxide capture capacity of the device and maintain the continuous operation of the device at the high-efficiency point. The controller sets the flow set value of the high-efficiency group as the first reference flow value, which corresponds to the upper limit of the high-efficiency region of the gas phase flow window of the sub-tower 400 (for example, 30000 Nm³ / h), so as to keep the carbon dioxide capture efficiency of the sub-tower 400 at a high level. In the embodiment, TA1 and TA2 are divided into the high-efficiency group.

[0088] The buffer group is an elastic unit for coping with load fluctuations. Under the rated operating condition, the buffer group maintains sufficient adjustment margin. The controller sets the flow set value of the buffer group as the second reference flow value, which is set at a flow level (for example, 22000 Nm³ / h) higher than the lower limit of the high-efficiency region but significantly lower than the upper limit of the high-efficiency region, so as to reserve sufficient adjustment space. When the load fluctuates, the controller absorbs the change by preferentially adjusting the load of the buffer group, so as to ensure that the sub-towers 400 of the high-efficiency group can stably operate at the high-efficiency point. In the embodiment, TA3 is divided into the buffer group.

[0089] The standby group. The hot standby can be realized by the following way: keeping the lean liquid and rich liquid cut-off valves of the sub-towers 400 of the standby group in the open state, allowing a small amount of liquid phase to circulate at a low speed in the tower to maintain the temperature, and at the same time keeping the tower system at a slight positive pressure (for example, 0.02 MPa gauge pressure) to prevent air from being sucked back. The controller isolates the standby group from the gas path by closing the inlet proportional regulating valve 500, which can be quickly enabled when needed. In the embodiment, TA4 is divided into the standby group.

[0090] When the device needs to stabilize the upstream converted gas volume at the rated load range (for example, 90000 Nm³ / h), the controller executes the following stable operation strategy:

[0091] State confirmation and load distribution: the controller confirms that the total gas volume upstream is in the rated value range, and then controls TA1 and TA2 of the high-efficiency group to operate at the first reference flow value (30000 Nm³ / h); controls TA3 of the buffer group to operate at the second reference flow value (22000 Nm³ / h for each sub-tower). At this time, the preset total processing gas volume of the device is 82000 Nm³ / h, and the difference of 8000 Nm³ / h is the adjustment capacity reserved by the controller for the buffer group to absorb the upward fluctuation of the load.

[0092] Pressure stabilization control: the controller ensures the pressure in the distributor 300 is stabilized at a preset value by adjusting the inlet regulating valve 200 on the total gas pipe 100.

[0093] Maintaining standby: the controller maintains the standby group (TA4) in a hot standby state, ensuring that it can be put into operation within a short preparation time (e.g., within ten minutes).

[0094] The controller pre-stores preset first opening threshold and second opening threshold, and is configured to perform the following dynamic adjustment strategy:

[0095] When receiving an upstream load change signal indicating an increase in load, preferentially increase the flow set value of the sub-tower 400 of the buffer group;

[0096] When the opening of the proportional regulating valve 500 of all sub-towers 400 in the buffer group reaches or exceeds the first opening threshold, it is determined that the buffer group has insufficient adjustment margin, and the flow set value of the sub-tower 400 of the high-efficiency group is increased again;

[0097] If the load continues to increase, the flow set values of the high-efficiency group and the buffer group all reach the first reference flow value, and the opening of the proportional regulating valve 500 of any sub-tower 400 in the buffer group and the high-efficiency group is above the preset first opening threshold, the controller activates one or more sub-towers 400 from the standby group and divides the corresponding sub-tower 400 into the buffer group;

[0098] When receiving an upstream load change signal indicating a decrease in load, preferentially decrease the flow set value of the sub-tower 400 of the buffer group;

[0099] When the opening of the proportional regulating valve 500 of the sub-tower 400 of the buffer group is reduced to below the preset second opening threshold, it is determined that the adjustment capacity of the buffer group is close to the lower limit, and the controller isolates the sub-tower 400 in the buffer group in turn and divides the corresponding isolated sub-tower 400 into the standby group.

[0100] The controller is configured to:

[0101] Based on real-time data collected by the sensor group, calculate the instantaneous operating state parameters of each sub-tower 400, including the opening of the proportional regulating valve 500;

[0102] Based on real-time data, the controller predicts the performance parameter trend of each sub-tower 400 in a future period of time through a built-in prediction model based on long short-term memory neural network;

[0103] Make a joint decision by combining the instantaneous operating state parameters and the predicted trend:

[0104] When the opening of the proportional regulating valve 500 does not reach the first opening threshold value, and the predicted trend indicates that the carbon dioxide concentration at the outlet will exceed the preset concentration alarm value at any time within the prediction time window, the preventive regulation is started, and the load distribution is adjusted in advance;

[0105] When the opening of the proportional regulating valve 500 reaches or exceeds the first opening threshold value, the regulation is carried out by the dynamic adjustment strategy.

[0106] The first opening threshold value is set in the range of 80% to 95%, and the second opening threshold value is set in the range of 15% to 30%.

[0107] In the embodiment of the present application, the controller pre-stores the first opening threshold value (such as 85%) and the second opening threshold value (such as 25%) for judging the upper limit and the lower limit of the regulation capacity of the sub-tower 400, respectively.

[0108] When the controller receives the upstream load increase signal (such as the total gas quantity set value is adjusted upward), the flow set value of each sub-tower 400 in the buffer group is first increased, and the newly added load is absorbed by increasing the opening of the proportional regulating valve 500.

[0109] If the opening of the proportional regulating valve 500 of all sub-towers 400 in the buffer group reaches or exceeds the first opening threshold value (85%), it is determined that the buffer group has insufficient regulation margin, at this time, the controller starts to increase the flow set value of the sub-tower 400 in the high-efficiency group, and utilizes the remaining regulation capacity.

[0110] If the load continues to increase, causing the flow set value of all running sub-towers 400 in the high-efficiency group and the buffer group to reach the first reference flow value (i.e. the upper limit of the high-efficiency zone), and the opening of the proportional regulating valve 500 of at least one sub-tower 400 still maintains above the first opening threshold value, the controller determines that the current running tower group has insufficient regulation capacity, at this time, the controller enables a sub-tower 400 from the standby group, opens the lean liquid and rich liquid shut-off valves of the sub-tower 400 to establish liquid circulation; slowly opens the gas proportional regulating valve 500, and the initial flow set value is set to the second reference flow value (such as 22000 Nm³ / h); the sub-tower 400 is divided into the buffer group, and the total load is redistributed so that the flow of all running sub-towers 400 returns to the high-efficiency zone.

[0111] When receiving the load reduction signal, the controller first reduces the flow set value of the sub-tower 400 in the buffer group, and reduces the opening of the proportional regulating valve 500.

[0112] When the opening of the proportional regulating valve 500 of all running sub-towers 400 in the buffer group is reduced to below the second opening threshold value (25%), it is determined that the regulation capacity is close to the lower limit. At this time, the controller starts the isolation program:

[0113] Select one buffer group sub-tower 400 (for example, the longest running time or relatively low efficiency), set its flow rate to zero, close its gas proportional valve 500 and lean liquid, rich liquid shut-off valve, and divide the sub-tower 400 into a standby group (default setting is cold standby, i.e. liquid path is closed and the tower is pressure maintained), while increasing the flow rate set value of the remaining running sub-tower 400 to maintain the operating point in the high efficiency region. If the load continues to decrease, the process can be repeated to isolate the buffer group sub-tower 400 one by one.

[0114] To improve the control foresight, the controller integrates a data-driven prediction model to predict future operation trends.

[0115] Prediction model functions:

[0116] The prediction model takes time series data collected by the sensor group as input, including real-time and historical values (such as data from the past 30 minutes) of the gas inlet flow rate, tower pressure, and outlet carbon dioxide concentration of each sub-tower 400. The model outputs the pressure drop change rate and outlet carbon dioxide concentration trend of each sub-tower 400 within a 5 to 15 minute prediction window (10 minutes in this embodiment).

[0117] Joint decision mechanism:

[0118] The controller combines the current proportional valve 500 opening and the predicted trend to perform the following two types of control:

[0119] Case one: preventive control is executed (valve position has a margin)

[0120] When the following two conditions are met:

[0121] Condition one: for sub-towers 400 whose predicted outlet carbon dioxide concentration will exceed the preset alarm value within the next 10 minutes, the corresponding proportional valve 500 opening has not yet reached the first opening threshold (85%);

[0122] Condition two: the prediction model shows that if the current operating state is maintained, the outlet carbon dioxide concentration of at least one sub-tower 400 will exceed the preset alarm value (such as 2.0 vol%) within the next 10 minutes;

[0123] The controller immediately starts preventive control, which uses a rolling horizon optimization algorithm to minimize the regeneration energy consumption of the regeneration tower (such as the minimum steam consumption of the reboiler), as the optimization objective, and all sub-tower 400 outlet carbon dioxide concentrations not exceeding the preset limit as a hard constraint, to calculate the optimal flow rate set value sequence of each sub-tower 400 within a certain period of time online, and gradually execute load redistribution.

[0124] Case two: regular dynamic adjustment is executed (valve position has reached the upper limit)

[0125] When the opening of any proportioning valve 500 of any sub-column 400 reaches or exceeds 85%, the controller no longer relies on the prediction model for optimal allocation, but directly executes the aforementioned dynamic adjustment strategy based on the opening threshold (such as enabling the standby column) to ensure the safe operation of the device.

[0126] Dynamic weighting of optimization objectives:

[0127] When performing preventive regulation, the controller dynamically adjusts the optimization priority according to the severity of the prediction results:

[0128] If the prediction shows that the carbon dioxide concentration at the outlet of all sub-columns 400 will be below the alarm value within the next 10 minutes, it is determined that the system is in a safe state, and the optimization algorithm prioritizes the lowest energy consumption;

[0129] If the prediction shows that any sub-column 400 will exceed the carbon dioxide concentration, it is determined that there is an environmental or process risk, and the optimization algorithm prioritizes ensuring that the concentration meets the standard, and then tries to reduce energy consumption as much as possible.

[0130] The performance parameter trend predicted by the prediction model includes the pressure drop change rate of the sub-column 400 or the change trend of the carbon dioxide concentration at the outlet, and the prediction time window is 5 to 15 minutes.

[0131] The prediction model is a data-driven model based on a long short-term memory neural network, which includes time series data of the inlet flow, tower pressure, and outlet carbon dioxide concentration of each sub-column 400 in the past 30 minutes as input, and the predicted value of the outlet carbon dioxide concentration of each sub-column 400 in the next 5 to 15 minutes as output.

[0132] The preventive regulation adopts a rolling horizon optimization algorithm, and the optimization algorithm calculates the optimal inlet flow distribution scheme of each sub-column 400 based on a pre-set mapping relationship between regeneration energy consumption and sub-column 400 load, and under the constraint that the outlet carbon dioxide concentration does not exceed the pre-set limit value. The mapping relationship is a function relationship of the regeneration energy consumption with respect to the inlet flow of the sub-column 400 and the corresponding rich liquid flow and carbon dioxide concentration, which is obtained by fitting historical operation data.

[0133] The controller is configured to adjust the weights of the optimization objectives and constraints according to the relationship between the predicted outlet carbon dioxide concentration and the pre-set concentration alarm value when performing preventive regulation: when the predicted outlet carbon dioxide concentration is lower than the pre-set concentration alarm value within the future prediction time window, the energy consumption optimization is prioritized;

[0134] When the predicted outlet carbon dioxide concentration will exceed the pre-set concentration alarm value at any time within the future prediction time window, the concentration constraint is prioritized and the energy consumption optimization is temporarily suspended.

[0135] In the embodiment of the present application, the prediction time window is set to:

[0136] If the prediction time window is short, the prediction result will not leave enough response time for the control system, making it difficult to effectively implement preventive control. If the prediction time window is long, the prediction accuracy will decrease due to the uncertainty factors such as component fluctuation and equipment aging in the chemical process. The prediction time window of the prediction model in the controller is set to 5 to 15 minutes, and the embodiment preferably is 10 minutes.

[0137] In the embodiment of the application, a data-driven prediction model based on a long short-term memory neural network is used.

[0138] The long short-term memory neural network includes an input layer, two LSTM hidden layers (the number of neurons is 64 and 32 respectively), and a fully connected output layer. The activation function uses ReLU, and the loss function uses mean square error.

[0139] The input of the long short-term memory neural network model includes the time series data of the gas flow, the pressure in the tower, and the carbon dioxide concentration at the gas outlet of each sub-tower 400 in the past 30 minutes. The output is the predicted value of the carbon dioxide concentration at the gas outlet of each sub-tower 400 in the future 5 to 15 minutes (preferably 10 minutes in the embodiment).

[0140] The long short-term memory neural network can be obtained through offline training: using historical operation data (including normal working conditions and disturbance working conditions) to construct a training set, using the actually measured future carbon dioxide concentration as a label, and using a back propagation algorithm to optimize network parameters. After the model training is completed, it is deployed in the controller and supports online fine-tuning to adapt to the slow changes in device performance (such as filler aging and amine degradation).

[0141] The historical operation data is derived from the actual operation records of the device for more than 3 months after normal operation, with a data sampling interval of 1 minute, covering the operation state under different load conditions. Data preprocessing includes outlier rejection, data normalization, and time series alignment.

[0142] It should be noted that although the long short-term memory neural network model is preferred in the embodiment, other model structures suitable for time series prediction (such as GRU, Transformer, or physical information neural network) can also be selected by those skilled in the art according to actual needs, as long as they can effectively predict the carbon dioxide concentration at the gas outlet based on the input data.

[0143] The specific form of the regeneration energy consumption mapping relationship is:

[0144] In the rolling horizon optimization algorithm, a regenerative energy consumption model is constructed as a function of the sub-tower 400 inlet flow rate, the corresponding rich liquid flow rate, and the carbon dioxide concentration in the rich liquid. The rich liquid flow rate can be obtained by a flow meter on the rich liquid header, and the carbon dioxide concentration in the rich liquid can be indirectly obtained by material balance calculation (e.g., based on the total amount of inlet carbon dioxide, the outlet carbon dioxide concentration, and the amine liquid circulation amount) or directly measured by an online rich liquid analyzer.

[0145] In this embodiment, the rolling horizon optimization uses a sequential quadratic programming algorithm for online solution, with an optimization horizon of 10 minutes and a control period of 30 seconds.

[0146] The mapping relationship of regenerative energy consumption is established by a multiple linear model, which expresses the regenerative energy consumption as a function of the sub-tower 400 inlet flow rate, the rich liquid flow rate, and the carbon dioxide concentration in the rich liquid. The specific coefficients in the model, including a constant term and three coefficient terms corresponding to the above operating variables, are determined by multiple linear regression analysis. This analysis is based on historical operation data sets collected after the system is put into normal operation for more than three months, and the least squares method is used for parameter fitting to determine a set of coefficients that can optimally reflect the influence of each variable on regenerative energy consumption. To ensure the prediction accuracy and reliability of the model, a goodness-of-fit verification is required after the model is established, with a requirement that the determination coefficient is not less than 0.85 to ensure that the model can be effectively used for subsequent optimization calculations.

[0147] In this embodiment, the mapping relationship is used to estimate the regenerative energy consumption under different load distribution schemes during optimization, thereby supporting decision-making with the goal of minimizing energy consumption.

[0148] The controller dynamically adjusts the target priority in the optimization process according to the severity of the prediction results, achieving intelligent balance between safety and economy.

[0149] For example, when the prediction results show that the outlet carbon dioxide concentration of all sub-towers 400 will be stably below the preset alarm value within the next 10 minutes, the controller determines that the system is in a safe state. At this time, the optimization algorithm takes the minimum regenerative energy consumption as the only optimization goal, and freely searches for the optimal load distribution scheme under the premise of meeting the concentration limit value.

[0150] When the prediction results show that the outlet carbon dioxide concentration of at least one sub-tower 400 will exceed the preset alarm value within the next 10 minutes, the controller determines that there is a risk of exceeding the standard. At this time, the optimization algorithm raises the concentration constraint to an unbreakable hard condition, prioritizing to ensure that all feasible solutions meet the concentration limit value. On this basis, the regenerative energy consumption is relatively lowest one is selected from all safe schemes as the execution instruction.

[0151] It should be noted that the connection mode of the lean-liquid and rich-liquid pipelines between the regenerator, the lean-liquid and rich-liquid heat exchanger and each sub-tower 400 is designed according to the conventional design in the art, which belongs to the prior art, and thus will not be described in detail herein;

[0152] The proportional regulating valve 500 arranged on the branch pipe 800 of the sub-tower 400 in the embodiment is a one-way valve, so that the shift gas in the sub-tower 400 is not easy to flow back to the distributor 300;

[0153] The total gas pipe, the regenerator (not shown in the figure) and the lean-liquid and rich-liquid heat exchanger (not shown in the figure) involved in the present application are all conventional basic equipment in the amine decarbonization process. The total gas pipe is used to transport the raw shift gas from the upstream section. The regenerator realizes the regeneration and circulation of the amine liquid by desorbing the carbon dioxide in the rich amine liquid through heating. The lean-liquid and rich-liquid heat exchanger is used to recover the heat carried by the rich amine liquid, and to reduce the overall energy consumption of the system by heat exchange with the cold lean amine liquid. The structure, function and basic process connection relationship between these devices are known in the prior art, and thus will not be described in detail herein.

[0154] The above is the preferred embodiment of the present application. It should be noted that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which should also be considered within the scope of protection of the present application.

Claims

1. A carbon dioxide capture device, comprising a main gas pipe, a regeneration tower, and a lean-rich liquid heat exchanger, wherein an inlet regulating valve is provided on the main gas pipe, characterized in that, Also includes: An absorption tower group consists of several sub-towers with identical structures. The distributor, connected to the main gas pipe via a pipeline, is used to receive the change gas from upstream, and the sub-tower is connected to the distributor via a branch pipe; The flow control valve assembly includes a proportional regulating valve installed on each of the branch pipes for independently controlling the flow rate of the change gas entering the corresponding sub-tower; The sensor group is used to collect data on the inlet flow rate, internal pressure, and carbon dioxide concentration at the outlet of each of the sub-towers in real time, as well as the pressure data inside the distributor. The controller, communicatively connected to the sensor group, the flow control valve group, and the intake regulating valve, is configured to: Based on the pressure data inside the distributor collected by the sensor group, the pressure inside the distributor is maintained at a preset pressure value by adjusting the opening of the air intake regulating valve. Based on the air intake flow data of each sub-tower collected by the sensor group, the flow rate of the changing gas entering each operating sub-tower is controlled at its preset flow rate setting value by independently adjusting the opening of the proportional regulating valve corresponding to the sub-tower. Based on the upstream load change signal and the tower pressure data and corresponding outlet carbon dioxide concentration data of the sub-tower collected by the sensor group, the number of sub-towers in operation and the corresponding inlet flow rate set value in the absorption tower group are dynamically adjusted so that the sub-towers in operation work in the high-efficiency zone of the corresponding gas phase flow window. The high-efficiency zone refers to the inlet flow rate range corresponding to the sub-tower under the condition that the carbon dioxide capture efficiency is not less than 95% and the pressure drop inside the tower does not exceed 30 kPa. The controller divides the absorption tower group into a high-efficiency group, a buffer group, and a standby group according to the control logic. The high-efficiency group consists of the minimum number of the sub-towers required to process the rated gas volume, and is used to continuously operate at the high-efficiency point under rated operating conditions; The buffer group, consisting of at least one of the sub-towers, is used to bear load changes to a limited extent during load fluctuations, so as to maintain the stable operation of the high-efficiency group under rated operating conditions; The standby group consists of at least one of the sub-towers and is in a cold standby or hot standby state. The controller has a first reference flow rate value and a second reference flow rate value stored in it. The first reference flow rate value corresponds to the upper limit of the high-efficiency zone of the gas phase flow window, and the second reference flow rate value corresponds to the lower limit of the high-efficiency zone of the gas phase flow window. The controller is configured to: when the total gas volume processed by the device corresponds to the rated load, control the high-efficiency group to operate at the first reference flow rate as the flow rate setpoint, control the buffer group to operate at the second reference flow rate as the flow rate setpoint, and maintain the standby group in standby state.

2. The carbon dioxide capture device according to claim 1, characterized in that, The controller has a preset first opening threshold and a second opening threshold, and is configured to execute the following dynamic adjustment strategy: When an upstream load change signal indicating an increase in load is received, the flow rate setting value of the sub-tower of the buffer group is increased first. When the opening degree of the proportional regulating valve of all sub-towers in the buffer group reaches or exceeds the first opening degree threshold, it is determined that the regulating margin of the buffer group is insufficient, and then the flow rate setting value of the sub-tower of the high efficiency group is increased. If the load continues to increase, causing the flow rate setpoints of both the high-efficiency group and the buffer group to reach the first reference flow rate value, and the opening degree of the proportional control valve of any of the sub-towers in the buffer group and the high-efficiency group is above the preset first opening degree threshold, the controller activates one or more sub-towers from the standby group and assigns the corresponding sub-tower to the buffer group. When an upstream load change signal indicating a reduction in load is received, the flow rate setting value of the sub-tower of the buffer group is preferentially reduced. When the opening degree of the proportional regulating valve of the sub-tower of the buffer group is reduced to below the preset second opening threshold, it is determined that the regulating capacity of the buffer group is close to the lower limit. The controller then isolates the sub-towers in the buffer group in sequence and assigns the corresponding isolated sub-towers to the standby group.

3. The carbon dioxide capture device according to claim 2, characterized in that, The controller is configured to: Based on the real-time data collected by the sensor group, the instantaneous operating status parameters of each sub-tower are calculated, including the opening degree of the proportional control valve. Based on the real-time data, the controller predicts the performance parameter trends of each sub-tower over a future period of time using a built-in prediction model based on a long short-term memory neural network, and outputs the predicted trends. A joint decision is made by combining the instantaneous operating state parameters with the predicted trend: When the opening degree of the proportional control valve does not reach the first opening degree threshold, and the predicted trend indicates that the carbon dioxide concentration at the outlet will exceed the preset concentration alarm value at any time within the predicted time window, preventive control is initiated to adjust the load distribution in advance. When the opening degree of the proportional control valve has reached or exceeded the first opening threshold, the dynamic adjustment strategy is used for regulation.

4. The carbon dioxide capture device according to claim 3, characterized in that, The first opening threshold is set in the range of 80% to 95%, and the second opening threshold is set in the range of 15% to 30%.

5. The carbon dioxide capture device according to claim 3, characterized in that, The predicted performance parameter trends include the rate of change of pressure drop in the sub-tower or the trend of change of carbon dioxide concentration at the outlet, with a prediction time window of 5 to 15 minutes.

6. The carbon dioxide capture device according to claim 3, characterized in that, The prediction model is a data-driven model based on a long short-term memory neural network. Its input includes time series data of the inlet flow rate, tower pressure and outlet carbon dioxide concentration of each sub-tower in the past 30 minutes, and the output is the predicted value of the outlet carbon dioxide concentration of each sub-tower in the next 5 to 15 minutes.

7. The carbon dioxide capture device according to claim 3, characterized in that, The preventive control adopts a rolling time-domain optimization algorithm. The optimization algorithm is based on a preset mapping relationship between regeneration energy consumption and sub-tower load, and calculates the optimal air intake flow distribution scheme for each sub-tower under the constraint that the outlet carbon dioxide concentration does not exceed the preset limit. The mapping relationship is a function relationship between regeneration energy consumption and sub-tower air intake flow, corresponding rich liquid flow and carbon dioxide concentration obtained by fitting historical operating data.

8. The carbon dioxide capture device according to claim 7, characterized in that, The controller is configured to adjust the weights of the optimization objective and constraints based on the relationship between the predicted outlet carbon dioxide concentration and the preset concentration alarm value when performing the preventative control: When the predicted export carbon dioxide concentration is lower than the preset concentration alarm value within the future prediction time window, energy consumption optimization will be given priority. If the predicted export carbon dioxide concentration exceeds the preset concentration alarm value at any time within the future prediction time window, then the concentration constraint will be prioritized and energy consumption optimization will be temporarily suspended.

Images