Efficient and energy-saving ship tail gas carbon capture system

Abstract

The invention provides an efficient and energy-saving ship tail gas carbon capture system which comprises a tail gas pretreatment module, a carbon capture module, a carbon separation and recovery module and an intelligent control unit. The intelligent control unit comprises a central processor, a data acquisition module, a communication module, an execution module and a data storage module. The system has a multi-module synergistic effect, from tail gas pretreatment to carbon capture and recovery, each module design aims at reducing energy consumption and improving efficiency, intelligence and physical rules are combined, key parameters are dynamically optimized through combination of a neural network and physical constraints, prediction precision and system performance are improved, advanced materials and methods are adopted, and the system is suitable for large-scale popularization and application. Through selection of adsorption materials (MOFs and amine modified zeolite) and use of a low-energy-consumption desorption method, energy consumption is reduced from the material level and the process level, and good balance of the system among capture efficiency, energy saving performance and environmental friendliness is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of ship exhaust gas treatment, and in particular to a highly efficient and energy-saving ship exhaust gas carbon capture system. Background Art

[0002] With the rapid development of the international shipping industry, ships, as important means of transportation, have played an important role in the increase of global greenhouse gas concentrations and the worsening of environmental pollution problems. Ship exhaust contains a large amount of carbon dioxide (CO2), particulate matter and other pollutants, which is one of the main sources of greenhouse effect and air quality degradation. Therefore, the development of efficient and energy-saving ship exhaust carbon capture technology to reduce CO2 emissions has important environmental significance and economic value.

[0003] Existing ship exhaust carbon capture technologies mainly include chemical absorption, physical adsorption and membrane separation, but there are still the following technical problems in practical application:

[0004] Insufficient exhaust gas pretreatment: Since the exhaust gas of ships contains particulate matter, high temperature and impurities, if it is not fully pretreated, it may affect the efficiency and service life of the subsequent capture device. Existing exhaust gas pretreatment devices are difficult to meet the dynamic adjustment and energy-saving requirements while ensuring high-efficiency filtration.

[0005] Unstable carbon capture efficiency: Existing adsorption technology exhibits unstable efficiency under complex ship operating conditions (such as fluctuations in exhaust gas temperature, flow rate and composition), and is difficult to adapt to the ever-changing exhaust gas characteristics during ship operation.

[0006] Lack of intelligent control: Existing carbon capture systems usually use simple mechanical control methods, which makes it difficult to optimize and control the dynamic changes of multiple parameters in real time, resulting in reduced system operating efficiency.

[0007] Insufficient integration of physical laws and machine learning: The control model of the traditional carbon capture system lacks the constraints of physical laws and relies on pure data-driven, which may lead to physical inconsistencies and inaccurate predictions and cannot efficiently guide the optimized operation of the system. Summary of the invention

[0008] The purpose of the present invention is to provide an efficient and energy-saving ship exhaust carbon capture system, which can significantly reduce the operating energy consumption of the ship exhaust carbon capture system, improve the capture efficiency and recovery purity, and realize intelligent and refined control, thereby providing an effective technical solution for the carbon neutrality goal in the ship field.

[0009] To achieve the above objectives, the present invention proposes the following technical solution: a highly efficient and energy-saving ship exhaust carbon capture system, comprising:

[0010] An exhaust gas pretreatment module, the exhaust gas pretreatment module comprising a high temperature filter and a condensing device, the high temperature filter is used to remove particulate matter and impurities in the exhaust gas, and the condensing device cools the exhaust gas to a temperature range suitable for treatment through a heat exchanger with an adjustable condensing temperature;

[0011] A carbon capture module, the carbon capture module comprising an adsorption tower and an adsorption layer plate, the adsorption layer plate being arranged in the adsorption tower, and the adsorption layer plate being provided with an adsorption material for capturing carbon dioxide;

[0012] A carbon separation and recovery module, comprising a desorption tower and a carbon dioxide compression unit. The desorption tower releases CO2 from the adsorption material using a low-energy vacuum desorption method or a temperature difference switching method, and the carbon dioxide compression unit compresses the captured CO2 into a high-density liquid state;

[0013] An intelligent control unit includes a central processing unit, a data acquisition module, a communication module, an execution module and a data storage module. The central processing unit is an industrial-grade embedded processor. The central processing unit is used to run a neural network model algorithm to process data in real time and output control instructions. The data acquisition module collects sensor data. The communication module is responsible for real-time communication between sensors, execution devices and the central processing unit. The execution module transmits control instructions to the execution device, which is composed of a multi-channel relay or electric valve control system.

[0014] Furthermore, in the present invention, the adsorption material is amine-modified zeolite or metal organic framework material, MOFs.

[0015] Furthermore, in the present invention, the communication module is one of CAN bus, Modbus or Ethernet.

[0016] Furthermore, in the present invention, the intelligent control unit runs a multi-task neural network with shared features, and the input vector x of the multi-task neural network with shared features includes the key features of all steps: x = [ΔP, PM, V, T in , T target , Q, W, C in , T abs , P abs , R regen , ...];

[0017] The shared feature multi-task neural network includes a shared feature extraction layer:

[0018] h=f shared (W shared ·x+b shared ), W shared , b sharedare the weight and bias of the shared network respectively, h: global feature vector;

[0019] The multi-task neural network with shared features performs the following tasks, where each task builds a branch network based on the shared features:

[0020] Exhaust cooling temperature prediction task;

[0021] CO2 adsorption efficiency prediction task;

[0022] Desorption pressure prediction task;

[0023] System timing prediction task.

[0024] Furthermore, in the present invention, the exhaust gas cooling temperature prediction task neural network formula is:

[0025] T out =f cooling (W cooling ·h+b cooling );

[0026] T out : Exhaust gas temperature after cooling, W cooling and b cooling are the weights and bias parameters of the cooling task network, respectively, h: shared input features, including the initial exhaust temperature T in , flow rate V, heat transfer efficiency parameters;

[0027] Constrained by physical laws, the cooling process follows the heat exchange equation:

[0028] Q=m·cp·(T in -T out )Q: heat exchange, m: exhaust gas mass flow, c p : Specific heat capacity of exhaust gas, T in and T out are the initial and final temperatures of the exhaust gas, respectively;

[0029] Add constraints to the loss function: L cooling =MSE(T out , T out，true )+λ1·|QQ true |;

[0030] MSE: Mean square error of cooling temperature prediction, Q true : The actual heat transfer calculated from the measured value, λ1: Physical constraint weight.

[0031] Furthermore, in the present invention, the CO2 adsorption efficiency prediction task neural network formula is:

[0032]

[0033] Adsorption efficiency, W adsorption and b adsorption are the weight and bias parameters of the adsorption efficiency prediction network, h: shared features, including the tower temperature T abs , flow rate Q, adsorption material state R adsorb ;

[0034] Constrained by physical laws, the adsorption process follows the Lagrangian adsorption isotherm:

[0035] The partial pressure of CO2 in the tail gas, k: equilibrium constant of the adsorption material;

[0036] Add constraints to the loss function:

[0037] MSE: mean square error of adsorption efficiency prediction, λ2: physical constraint weight.

[0038] Furthermore, in the present invention, the neural network formula for the desorption pressure prediction task is:

[0039] P regen =f desorption (W desorption ·h+b desorption );

[0040] P regen : desorption pressure, W desorption and b desorption are the weight and bias parameters of the desorption pressure prediction network, respectively, h: shared features, including the desorption temperature T regen , adsorbent state R adsorb ;

[0041] Constraints of physical laws: The desorption process conforms to the gas state equation: P regen V = n R T regen ;

[0042] V: fixed volume of the desorption process, n: number of CO2 molecules adsorbed, R: ideal gas constant, T regen : Desorption temperature, add constraints to the loss function:

[0043] L desorption =MSE(P regen , P regen，true )+λ3·|P regen ·Vn·R·T regen |;

[0044] MSE: mean square error of desorption pressure prediction, λ3: physical constraint weight.

[0045] Furthermore, in the present invention, the neural network formula for the system timing prediction task is: t =LSTM(x t ,h t-1 , c t-1 );

[0046] h t : The state vector of the current time step, x t : Input characteristics of time step t, including flow rate V, temperature T, equipment state parameters, h t-1 and c t-1 are the hidden layer and cell states of the previous time step respectively;

[0047] Physical law constraints: system energy consumption E and flow rate V, temperature T and adsorption efficiency There is a relationship: η: system energy conversion efficiency;

[0048] Add constraints to the loss function:

[0049] λ4: physical constraint weight;

[0050] The final total loss function is the weighted sum of multiple subtasks, with dynamic physical law constraints added: L i : The loss function for each task, α i : Dynamic weight, adjusted with system performance and priority. The physical law constraint term is an important part of the loss function of each subtask, which is used to improve the physical consistency and prediction accuracy of the model.

[0051] Furthermore, in the present invention, the data acquisition module includes:

[0052] A first temperature sensor, which is installed before the tail gas enters the condensing device and at the outlet of the condensing device. The first temperature sensor monitors the initial temperature and the temperature after cooling of the tail gas to ensure that the tail gas is in a suitable temperature range when entering the adsorption tower;

[0053] A first gas flow rate sensor, which is installed before the exhaust gas enters the condensing device and at the outlet of the condensing device, and is used to measure the exhaust gas flow rate;

[0054] A particle sensor is installed after the high-temperature filter to detect the particle content in the exhaust gas and evaluate the filtering effect;

[0055] The second temperature sensor is installed at each layer of the adsorption layer plate inside the adsorption tower to monitor the temperature distribution inside the adsorption tower to ensure that the adsorption material operates within the optimal working temperature range;

[0056] A second gas flow rate sensor, which is installed at the inlet and outlet of the adsorption tower to measure the tail gas flow rate and provide a reference for predicting the adsorption efficiency;

[0057] A first gas composition sensor, which is installed at the entrance of the adsorption tower, between each adsorption layer plate, and at the exit, and is used to monitor the change of CO2 concentration and calculate the adsorption efficiency;

[0058] A third temperature sensor is installed inside and at the outlet of the desorption tower to monitor the temperature change during the desorption process and optimize the desorption conditions;

[0059] A pressure sensor is installed inside the desorption tower, at the outlet, and before the compression unit to monitor the desorption pressure and guide the system to optimize operation;

[0060] A second gas composition sensor, which is installed at the outlet of the analysis tower and is used to monitor the purity of CO2 to ensure the quality of the output gas;

[0061] Energy consumption sensors are installed at the power interfaces of the cooling device, adsorption tower, analytical tower and compression unit to record the energy consumption of each module in real time and optimize the system operation efficiency.

[0062] Furthermore, in the present invention, the execution module includes an electric valve, a multi-channel relay, a regulating heat exchanger and a pressure regulator. The electric valve controls the exhaust gas flow and direction. The electric valve is arranged before and after the condensing device, at the inlet and outlet of the adsorption tower, and at the outlet of the analysis tower. The multi-channel relay controls the switches of the cooling device, the heater and the compression unit. The multi-channel relay is installed at the power interface of each device. The regulating heat exchanger adjusts the cooling efficiency of the condensing device and is installed inside the condensing device. The force regulator controls the pressure change in the analysis tower and is installed inside the analysis tower.

[0063] Beneficial effects: The technical solution of this application has the following technical effects:

[0064] The present invention has a multi-module synergistic effect. From tail gas pretreatment to carbon capture and recovery, each module is designed to reduce energy consumption and improve efficiency. It combines intelligence with physical laws, dynamically optimizes key parameters through the combination of neural networks and physical constraints, improves prediction accuracy and system performance, and adopts advanced materials and methods: selection of adsorption materials (MOFs, amine-modified zeolites) and use of low-energy desorption methods to reduce energy consumption from the material and process levels. The present invention ensures a good balance between capture efficiency, energy saving and environmental friendliness of the system.

[0065] It should be appreciated that all combinations of the foregoing concepts, as well as additional concepts described in greater detail below, may be considered to be part of the inventive subject matter of the present disclosure, provided such concepts are not mutually inconsistent.

[0066] The foregoing and other aspects, embodiments and features of the present invention can be more fully understood from the following description in conjunction with the accompanying drawings. Other additional aspects of the present invention, such as the features and / or beneficial effects of the exemplary embodiments, will be apparent from the following description or learned from the practice of the specific embodiments according to the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component shown in various figures may be represented by the same reference numeral. For clarity, not every component is labeled in every figure. Embodiments of various aspects of the present invention will now be described by way of example and with reference to the accompanying drawings, in which:

[0068] Figure 1 It is a schematic diagram of the structure of the present invention.

[0069] Figure 2 It is a schematic diagram of the local structure of the present invention.

[0070] In the figure, the meanings of the various reference numerals are as follows: 1. hull; 2. cabin; 3. exhaust carbon capture system; 301. high-temperature filter; 302. condensation device; 303. adsorption tower; 304. analysis tower; 305. carbon dioxide compression unit; 306. storage tank. DETAILED DESCRIPTION

[0071] In order to better understand the technical content of the present invention, specific embodiments are cited and described as follows in conjunction with the accompanying drawings. Various aspects of the present invention are described in this disclosure with reference to the accompanying drawings, in which many illustrative embodiments are shown. The embodiments of the present disclosure are not necessarily defined to include all aspects of the present invention. It should be understood that the various concepts and embodiments introduced above, as well as those described in more detail below, can be implemented in any of many ways, because the concepts and embodiments disclosed in the present invention are not limited to any implementation. In addition, some aspects disclosed in the present invention can be used alone, or in any appropriate combination with other aspects disclosed in the present invention.

[0072] As shown in Figures 1-2, this embodiment provides an efficient and energy-saving ship exhaust carbon capture system, which is installed on the hull 1 of the ship. The hull 1 includes a cabin 2 and an exhaust carbon capture system 3. The exhaust carbon capture system 3 includes an exhaust pretreatment module, a carbon capture module, a carbon separation and recovery module, and an intelligent control unit. The exhaust pretreatment module includes a high-temperature filter 301 and a condensing device 302. The high-temperature filter 301 is used to remove particulate matter and impurities in the exhaust gas. The condensing device 302 cools the exhaust gas to a temperature range suitable for treatment through a heat exchanger with an adjustable condensing temperature; the high-temperature filter 301 removes particulate matter and impurities in the exhaust gas, protects subsequent equipment from damage, and improves the overall efficiency of the system. The condensing device 302 cools the exhaust gas to a temperature range suitable for treatment, providing optimal operating conditions for the adsorption process of the carbon capture module. The high-temperature filter 301 improves the stability and reliability of the system operation, reduces the frequency of equipment maintenance, and the cooling temperature of the condensing device 302 is adjustable, which can flexibly adapt to different exhaust conditions, avoid excessive temperature affecting the performance of the adsorption material, and reduce the energy consumption of subsequent treatment. Efficient heat recovery and exhaust gas cooling are achieved through the heat exchanger, reducing the burden on subsequent adsorption modules and overall energy consumption.

[0073] The carbon capture module includes an adsorption tower 303 and an adsorption layer plate. The adsorption layer plate is arranged in the adsorption tower 303. The adsorption layer plate is provided with an adsorption material for capturing carbon dioxide. The adsorption material is amine-modified zeolite or metal organic framework material, MOFs; the adsorption tower 303 and the adsorption layer plate provide a place for capturing and fixing carbon dioxide. The adsorption material (amine-modified zeolite or MOFs) provides efficient CO2 adsorption performance and improves the capture efficiency. The structural design of the adsorption layer plate allows uniform gas flow to ensure adsorption efficiency. Advanced adsorption materials have higher selectivity and adsorption capacity, reducing the energy consumption required per unit processing volume. The high adsorption efficiency reduces the need for recycling untreated carbon dioxide gas, thereby reducing energy consumption.

[0074] The carbon separation and recovery module includes an analysis tower 304 and a carbon dioxide compression unit 305. The analysis tower 304 uses a low-energy vacuum desorption method or a temperature difference switching method to release CO2 from the adsorption material. The adsorption layer in the adsorption tower 303 is lifted to the analysis tower 304 by a lifting device for analysis. The carbon dioxide compression unit 305 compresses the captured CO2 into a high-density liquid state and stores it in a storage tank 306. The carbon dioxide compression unit 305 can use a screw compressor. The screw compressor uses a pair of intermeshing screws to rotate so that the gas is gradually compressed. Its compression process is stable and pulsation-free. It is simple to maintain and has a long service life. It can handle wet gas and is suitable for carbon dioxide compression.

[0075] The analysis tower 304 releases carbon dioxide from the adsorbent material using a vacuum desorption method or a temperature difference switching method. The carbon dioxide compression unit 305 compresses the CO2 obtained by analysis into a high-density liquid state for easy storage and transportation. The vacuum desorption method and the temperature difference switching method have low energy consumption, reducing the energy input of the desorption process. The compression unit design is optimized to achieve the liquefaction of CO2 quickly and efficiently. Vacuum desorption or temperature difference switching reduces the heat energy demand of the traditional heating desorption method, greatly reducing the energy consumption of the desorption process.

[0076] The intelligent control unit includes a central processing unit, a data acquisition module, a communication module, an execution module and a data storage module. The central processing unit is an industrial-grade embedded processor. The central processing unit is used to run the neural network model algorithm to process data in real time and output control instructions. The data acquisition module collects sensor data. The communication module is responsible for real-time communication between sensors, actuators and central processing units. The communication module is a CAN bus. The execution module transmits control instructions to the actuator. The control system consists of a multi-channel relay or electric valve control system. The intelligent control unit provides optimization control based on the neural network algorithm and dynamically adjusts system parameters (such as cooling temperature, desorption pressure, adsorption efficiency, etc.). Collect multi-sensor data and provide real-time feedback to improve operating efficiency. The neural network algorithm dynamically predicts and adjusts the status of each module to ensure that the system is always in the best working state. Support multiple communication protocols (CAN bus, Modbus, Ethernet) to enhance system compatibility and scalability. Intelligent prediction and dynamic optimization reduce unnecessary energy consumption of the system, such as avoiding energy waste caused by overcooling or overheating, and optimizing the start and stop sequence of equipment to improve overall energy efficiency.

[0077] The data acquisition module monitors key parameters (temperature, flow rate, pressure, gas composition, etc.) in real time, providing reliable data support for the intelligent control unit. The monitoring is comprehensive, covering key performance indicators of all links such as condensation, adsorption, and analysis. The design of multi-point monitoring improves the safety and accuracy of system operation. Energy saving: Data-driven optimization strategy ensures reasonable distribution of energy consumption and reduces energy waste.

[0078] The execution module achieves precise control of flow, direction, cooling efficiency and pressure through electric valves, multi-channel relays, modulating heat exchangers and pressure regulators. All components work in coordination and respond quickly to ensure dynamic adaptability under different working conditions. The modulating heat exchanger further improves the thermal energy utilization rate of the condensation process. Accurately control the operating conditions of the equipment to avoid ineffective energy consumption and improve the overall energy efficiency of the system.

[0079] In summary, energy recovery is optimized through heat exchange, reducing dependence on additional cooling equipment. Amine-modified zeolite and MOFs materials have high adsorption performance, reducing energy consumption per unit capture volume. Vacuum desorption and temperature difference switching methods significantly reduce the energy demand of the desorption process. Dynamic tuning and energy-saving operation are achieved by using physical law constraints and multi-task neural network models. Real-time data collected by sensors provide a reliable basis for energy efficiency optimization. Physical law constraints are added to the neural network to improve the prediction accuracy of the model and avoid energy waste caused by misoperation. Through the synergy of the above modules and intelligent control optimization, the system minimizes energy consumption in ship exhaust carbon capture, meeting the goal of high efficiency and energy saving.

[0080] Furthermore, in this embodiment, the intelligent control unit runs a multi-task neural network with shared features, wherein the multi-task neural network with shared features can make the expression of the model more accurate, while reducing the computational complexity, and improving real-time performance and system performance. By introducing multi-task learning, it is not necessary to establish a neural network for each step separately, but to build a multi-task neural network with shared features. The first few layers of the network share the global features of the exhaust treatment system (such as flow rate, temperature, particulate matter concentration, etc.), and the subsequent branches implement different tasks (such as temperature control, adsorption efficiency prediction, etc.), reduce model parameters, improve data utilization efficiency, and enhance the ability to collaborate between tasks.

[0081] This embodiment also uses constraints based on physical laws, and incorporates the physical laws of the system (such as thermodynamic equations and gas dynamics equations) into the neural network loss function to improve the physical consistency of the model, reduce dependence on experimental data, and improve the generalization ability of the model. At the same time, the activation function and the loss function are optimized. The activation function uses the Swish or Mish function instead of the traditional ReLU to improve the nonlinear fitting ability of the model. The loss function uses a dynamic weighted loss function, focusing on optimizing the steps with low capture efficiency or high energy consumption.

[0082] This embodiment further introduces timing optimization and memory mechanism, and uses long short-term memory (LSTM) to process the dynamic timing characteristics of the system, such as the decay of adsorption material performance over time. It can capture the dynamic changes of the system and improve long-term optimization capabilities. Regularization and model compression are added. Regularization: Reduce overfitting through L1 / L2 regularization. Model compression: Introduce pruning or quantization technology to reduce computational complexity and facilitate real-time deployment.

[0083] The details are as follows:

[0084] The input vector x of the multi-task neural network with shared features includes the key features of all steps: x = [ΔP, PM, V, T in , T target , Q, W, C in , T abs , P abs , R regen , ...], ΔP means: pressure difference in the system (unit: kPa). Function: reflects the resistance of gas flow in the system, used to evaluate the operating performance of the adsorption tower and desorption tower. PM means: particulate matter concentration (unit: mg / m 3 ). Function: Indicates the level of particulate matter pollution in the exhaust gas, used to control and optimize the efficiency of the filter. V Meaning: Exhaust flow rate (unit: m / s). Function: Affects the exhaust cooling efficiency and adsorption efficiency, and is a key parameter for dynamic optimization. T in Meaning: The initial temperature of the tail gas entering the system (unit: °C). Function: As the main input for cooling temperature prediction, it ensures cooling to the optimal adsorption temperature range. target Meaning: Target temperature of exhaust gas cooling (unit: °C). Function: Indicates the temperature that the exhaust gas needs to reach after cooling, used to control the output of the cooling device. Q Meaning: Exhaust gas flow rate (unit: m 3 / h). Function: Affects the working efficiency of the adsorption tower and cooling device, and is a key parameter in the adsorption and desorption process. W Meaning: System energy consumption (unit: kW). Function: Real-time reflection of the energy usage of the system, used for energy recovery and optimization. C in Meaning: CO2 concentration entering the adsorption tower (unit: %). Function: The main input for predicting CO2 adsorption efficiency and evaluating the carbon dioxide content in the tail gas. abs Meaning: The internal temperature of the adsorption tower (unit: °C). Function: Directly affects the capture efficiency of the adsorption material and is an important input for dynamic regulation. abs Meaning: The internal pressure of the adsorption tower (unit: kPa). Function: Together with the temperature, it determines the performance of CO2 adsorption and optimizes the adsorption efficiency. regen Meaning: Desorption material regeneration rate (unit: kg / s). Function: Indicates the regeneration speed of the adsorbent material, used to optimize the operating parameters of the desorption tower. It also includes other parameter data required by the neural network, which are not listed here one by one. Just set the corresponding sensors or sensing devices as needed. These features provide rich input information for the multi-task neural network to support the efficient operation and optimization of tasks such as cooling temperature prediction, CO2 adsorption efficiency prediction, desorption pressure prediction and system timing prediction.

[0085] The shared feature multi-task neural network includes a shared feature extraction layer:

[0086] h=f shared (W shared ·x+b shared ), W shared , b shared are the weight and bias of the shared network respectively, h: global feature vector;

[0087] The multi-task neural network with shared features performs the following tasks, where each task builds a branch network based on the shared features:

[0088] Exhaust cooling temperature prediction task;

[0089] CO2 adsorption efficiency prediction task;

[0090] Desorption pressure prediction task;

[0091] System timing prediction task.

[0092] Specifically, the neural network formula for the exhaust gas cooling temperature prediction task is:

[0093] T out =f cooling (W cooling ·h+b cooling );

[0094] T out : Exhaust gas temperature after cooling, W cooling and b cooling are the weights and bias parameters of the cooling task network, respectively, h: shared input features, including the initial exhaust temperature T in , flow rate V, heat transfer efficiency parameters;

[0095] Constrained by physical laws, the cooling process follows the heat exchange equation:

[0096] Q=m·c p ·(T in -T out )Q: heat transfer, m: exhaust gas mass flow, c p : Specific heat capacity of exhaust gas, T in and T out are the initial and final temperatures of the exhaust gas, respectively;

[0097] Add constraints to the loss function: L cooling =MSE(T out , T out，true )+λ1·|QQ true |;

[0098] MSE: Mean square error of cooling temperature prediction, Q true : The actual heat transfer calculated from the measured value. λ1: Physical constraint weight.

[0099] The exhaust gas cooling temperature prediction task can optimize the cooling process and predict the exhaust gas temperature T after cooling. out To ensure that the tail gas is in the optimal operating temperature range (40-70°C) of the adsorption material when entering the adsorption tower 303. It can prevent equipment damage and control the temperature to avoid damage to the adsorption material and cooling equipment caused by excessively high or low temperatures. It can dynamically adjust the heat exchanger parameters, and guide the cooling system to dynamically adjust the heat exchange efficiency by predicting the temperature in real time to ensure efficient operation.

[0100] Therefore, by accurately predicting Tout, we can avoid overcooling and reduce the energy consumption of the cooling device. We can improve the adsorption efficiency, cool to the optimal temperature range, and enhance the CO2 capture capacity of the adsorbent material. The advantages of physical law constraints, adding the heat exchange equation Q = m·c p ·(T in -T out ) to ensure that the predicted temperature meets the actual physical conditions and improve the reliability of the prediction. Avoid the model from producing output values ​​that do not conform to physical meaning.

[0101] Specifically, the neural network formula for the CO2 adsorption efficiency prediction task is:

[0102]

[0103] Adsorption efficiency, W adsorption and b adsorption are the weight and bias parameters of the adsorption efficiency prediction network, h: shared features, including the tower temperature T abs , flow rate Q, adsorption material state R adsorb ;

[0104] Constrained by physical laws, the adsorption process follows the Lagrangian adsorption isotherm:

[0105] The partial pressure of CO2 in the tail gas, k: equilibrium constant of the adsorption material;

[0106] Add constraints to the loss function:

[0107] MSE: mean square error of adsorption efficiency prediction, λ2: physical constraint weight.

[0108] The CO2 adsorption efficiency prediction task can predict the adsorption performance and calculate the CO2 capture efficiency of the adsorption material in the adsorption tower 303. It can dynamically optimize the parameters of the adsorption tower 303 and adjust the working temperature, pressure and gas flow rate of the adsorption tower 303 in real time according to the efficiency prediction results. It can guide the replacement or regeneration of the adsorption material. The adsorption efficiency prediction can determine the working state of the adsorption material and decide whether it needs to be replaced or regenerated.

[0109] Therefore, the model has high capture efficiency by accurately predicting Guide the multi-stage adsorption tower 303 to perform fine control to increase the CO2 capture efficiency to more than 90%. Extend the life of the adsorption material, avoid material overload operation, and extend the use cycle of the adsorbent. The advantage of physical law constraints is that the Lagrangian adsorption isotherm formula Match the prediction results with the material properties to avoid predictions beyond the adsorbent capacity. Improve the model generalization ability and reduce the dependence on data size.

[0110] Specifically, the neural network formula for the desorption pressure prediction task is:

[0111] P regen =f desorption (W desorption ·h+b desorption );

[0112] P regen : desorption pressure, W desorption and b desorption are the weight and bias parameters of the desorption pressure prediction network, respectively, h: shared features, including the desorption temperature T regen , adsorbent state R adsorb ;

[0113] Constraints of physical laws: The desorption process conforms to the gas state equation: P regen V = n R T regen ;

[0114] V: fixed volume of the desorption process, n: number of CO2 molecules adsorbed, R: ideal gas constant. regen : Desorption temperature, add constraints to the loss function:

[0115] L desorption =MSE(P regen , P regen，true )+λ3·|P regen ·Vn·R·T regen |;

[0116] MSE: mean square error of desorption pressure prediction. λ3: physical constraint weight.

[0117] The desorption pressure prediction task can guide the desorption process and predict the pressure P required in the desorption process. regen, ensuring that CO2 can be efficiently released from the adsorbent material. It can reduce energy consumption, optimize desorption conditions, and reduce energy waste during desorption. It can enhance regeneration efficiency, and predicting pressure can help achieve complete regeneration of the adsorbent material and improve material utilization.

[0118] Therefore, the equipment can be operated with low energy consumption. By accurately predicting the desorption pressure, unnecessary pressure increase or pressure reduction operations can be reduced, thereby reducing the energy consumption of the regeneration device. The regeneration efficiency is high. After the desorption conditions are optimized, the adsorbed CO2 can be completely released to ensure the regeneration capacity of the adsorption material. The advantages of physical law constraints, using the gas state equation P regen V = n R T regen , ensuring that the desorption pressure is consistent with physical parameters such as system temperature, volume, and adsorbent saturation to avoid invalid predictions. Combining physical formulas makes the network's prediction of desorption conditions more accurate and reliable.

[0119] Specifically, the neural network formula for the system timing prediction task is: t =LSTM(x t ,h t-1 , c t-1 );

[0120] h t : The state vector of the current time step, x t : Input characteristics of time step t, including flow rate V, temperature T, equipment state parameters, h t-1 and c t-1 are the hidden layer and cell states of the previous time step respectively;

[0121] Physical law constraints: system energy consumption E and flow rate V, temperature T and adsorption efficiency There is a relationship: η: system energy conversion efficiency;

[0122] Add constraints to the loss function:

[0123] λ4: physical constraint weight;

[0124] The final total loss function is the weighted sum of multiple subtasks, with dynamic physical law constraints added: L i : The loss function for each task, α i : Dynamic weight, adjusted with system performance and priority. The physical law constraint term is an important part of the loss function of each subtask, which is used to improve the physical consistency and prediction accuracy of the model.

[0125] The system timing prediction task can monitor the system status in real time and predict the dynamic changes in the system operation process, including exhaust gas flow rate, temperature fluctuations, equipment performance, etc. It can perform preventive maintenance and discover potential faults in advance by analyzing timing data to avoid equipment downtime. It can optimize the overall system performance and dynamically adjust the operating parameters of each module based on the results of timing prediction to achieve coordinated optimization of the entire system.

[0126] Therefore, it can improve system stability and avoid operation interruptions caused by sudden parameter anomalies by predicting the future state of the system. It helps to intelligentize energy management, optimize energy distribution and reduce overall system energy consumption by combining real-time data. Advantages of physical law constraints: As a constraint, it guides the model to output prediction results that comply with energy conservation. It strengthens the model's ability to understand complex time series data and avoids generating unreasonable time series features.

[0127] In summary, the network structure of this embodiment can mine the complex relationship between input features and has strong adaptability. Multi-task sharing feature h improves computing efficiency and avoids redundant calculations. Introducing physical laws into the loss function not only improves the accuracy of model predictions, but also enhances the physical consistency of the model. Reducing dependence on large-scale training data helps to deal with the problem of data scarcity. By guiding the network to follow real physical phenomena, the model outputs unreasonable prediction results. The constraints in the loss function are adjusted through dynamic weights (λ i ) regulation, the priority of each task can be adjusted according to the actual operation situation. Combined with the constraints of physical laws, the model's adaptability to complex working conditions is improved, especially in terms of energy optimization and system operation efficiency.

[0128] The system works as follows:

[0129] The exhaust gas first passes through the high temperature filter 301. The high temperature filter 301 uses high temperature resistant materials to intercept particulate matter and impurities in the exhaust gas, ensuring that the exhaust gas entering the subsequent processing equipment has a high degree of purity, thereby improving the system operation efficiency and reducing equipment wear.

[0130] The high-temperature filtered tail gas flows into the condensation device 302. The condensation device 302 cools the tail gas to a temperature range suitable for treatment through a heat exchanger with an adjustable condensation temperature. The cooling process uses a heat exchange medium to exchange heat with the tail gas, and by adjusting the cooling efficiency of the heat exchanger, the tail gas temperature reaches the optimal working range of the adsorption tower 303.

[0131] The cooled tail gas enters the adsorption tower 303, where multiple layers of adsorption plates are arranged. The adsorption plates are coated with amine-modified zeolite or metal organic framework materials (MOFs). The adsorption material captures carbon dioxide (CO2) molecules in the tail gas by combining chemical adsorption and physical adsorption: Chemical adsorption: Amine groups form stable chemical bonds with CO2, which has a high adsorption efficiency. Physical adsorption: The high specific surface area of ​​the adsorption material increases the ability to capture CO2 molecules.

[0132] The temperature, flow rate and gas composition sensors arranged inside the adsorption tower 303 monitor the temperature distribution, exhaust gas flow rate and CO2 concentration changes in the tower in real time to ensure the stability of the adsorption efficiency and the optimal working state of the adsorption material.

[0133] After the adsorption material is saturated, it is transported to the desorption tower 304 by the lifting equipment. The desorption tower 304 achieves CO2 desorption through the following two low-energy consumption methods: vacuum desorption method, which reduces the pressure in the tower, reduces the adsorption force, and releases CO2 from the surface of the adsorption material. The temperature difference switching method uses a heating device to increase the temperature in the tower, break the chemical bonds, and release the adsorbed CO2. The temperature and pressure of the desorption tower 304 are monitored and optimized in real time through built-in sensors to ensure that the desorption process is efficient and stable.

[0134] The CO2 gas released after analysis is transported to the compression unit through a pipeline. The compression unit compresses the CO2 into a high-density liquid state for easy storage or transportation. The whole process has low energy consumption and ensures energy saving.

[0135] The intelligent control unit coordinates and optimizes the control of each module of the system. Its working principle is as follows:

[0136] The data acquisition module integrates a variety of sensors (temperature, flow rate, pressure, composition, energy consumption, etc.) to collect physical parameters, equipment status and energy consumption data of tail gas in real time. The intelligent control unit runs a multi-task neural network model with shared features to predict the key parameters of the system in real time based on the data collected by the sensors: based on the initial temperature, flow rate and heat exchange efficiency, the temperature of the tail gas after cooling is predicted to ensure the optimal temperature entering the adsorption tower 303. Based on the temperature, flow rate and adsorption material state in the tower, the adsorption efficiency is predicted, and the operating state of the adsorption tower 303 is adjusted in time. According to the desorption temperature and adsorbent state, the desorption pressure is predicted to optimize the operating conditions of the analytical tower 304. The system energy consumption is predicted in time series, and the energy consumption distribution is adjusted based on the flow rate, temperature and adsorption efficiency to improve the overall energy saving effect. The neural network model combines physical law constraints (such as heat exchange equations, Lagrangian adsorption isotherms, gas state equations, etc.) to achieve high-precision predictions and improve control accuracy and physical consistency.

[0137] The intelligent control unit is connected to the execution device of each module through the communication module, and sends control instructions to adjust the operating parameters. For example: adjust the heat exchange efficiency of the condensing device 302. Control the inlet and outlet valve switches of the adsorption tower 303 and the analytical tower 304. Start or stop the cooling device, heater and compression unit. Specifically, the data acquisition module collects the initial temperature (before entering the condensing device) and the temperature after cooling (condensing device outlet) of the exhaust gas in real time through the first temperature sensor. The intelligent control unit runs the neural network model to predict the temperature of the exhaust gas after cooling based on the initial temperature, flow rate and heat exchange efficiency of the exhaust gas. If the cooling temperature deviates from the optimal value, the intelligent control unit sends instructions to the regulating heat exchanger through the communication module. After receiving the instruction, the regulating heat exchanger realizes dynamic adjustment of the heat exchange efficiency by changing the flow rate or temperature of the cooling medium in the heat exchanger. The updated cooling temperature is collected again by the sensor to form a closed-loop control.

[0138] The second temperature sensor monitors the internal temperature of the adsorption tower, the second gas flow rate sensor monitors the inlet and outlet flow rates of the adsorption tower, and the first gas composition sensor monitors the change in CO2 concentration. The neural network model predicts the adsorption efficiency based on the temperature, flow rate and state of the adsorption material in the tower. If the adsorption efficiency decreases, the intelligent control unit sends a command to the electric valve of the adsorption tower through the communication module. The electric valve at the inlet of the adsorption tower adjusts the exhaust gas flow to adapt to the state of the adsorption material, and the outlet electric valve ensures that the pressure in the tower remains stable. After receiving the command, the electric valve adjusts the opening to optimize the flow distribution. If the temperature in the adsorption tower deviates from the optimal range, the cooling device can be started or stopped to adjust the environment in the tower.

[0139] The third temperature sensor monitors the temperature inside and at the outlet of the analysis tower, the pressure sensor monitors the desorption pressure, and the second gas composition sensor monitors the CO2 purity. The neural network model combines the desorption temperature and the adsorbent state to predict the desorption pressure. If the desorption pressure is too high or too low, the intelligent control unit sends instructions to the pressure regulator and the electric valve inside the analysis tower through the communication module. The pressure regulator adjusts the pressure inside the analysis tower according to the instructions, and the electric valve controls the flow and direction of CO2 in the analysis tower. If the temperature is not enough to achieve efficient desorption, start the heater and adjust the power. If the pressure is too high, the pressure regulator releases excess gas to avoid abnormal operation of the analysis tower.

[0140] The pressure sensor monitors the gas pressure entering the compression unit, and the second gas composition sensor monitors the CO2 purity. If the compression unit detects that the gas pressure is too low, the intelligent control unit sends a command to start or increase the power of the desorption tower heater. If the gas pressure is moderate, the intelligent control unit sends a command to start the compression unit. The multi-channel relay controls the operating status of the compression unit, including power on and off and power adjustment. If the energy consumption of the compression unit is too high, the intelligent control unit reduces the desorption speed of the desorption tower to relieve the load.

[0141] Energy consumption sensors record the real-time energy consumption of the condensing unit, adsorption tower, analytical tower and compression unit. The neural network model predicts the system energy consumption and optimizes energy distribution based on flow rate, temperature and adsorption efficiency. If the energy consumption of a module exceeds the preset range, the intelligent control unit sends instructions through the communication module to adjust the operating parameters of the relevant equipment. For example, the power of the heater or compression unit is reduced through a multi-channel relay to avoid excessive overall energy consumption.

[0142] The intelligent control unit adds physical law constraints to the loss function of the model to ensure that the prediction results conform to the actual physical principles, thereby avoiding system instability caused by data noise or modeling errors. The dynamic weight mechanism of the multi-task neural network enables the system to adjust the priority of each task according to the actual operation situation, such as optimizing adsorption efficiency at high load and reducing energy consumption at low load. The data storage module records historical operation data to provide a basis for subsequent maintenance, optimization and fault diagnosis.

[0143] The entire system forms a closed-loop control, adjusting the operating parameters of each module in real time. The operating status monitored by the data acquisition module is used as input and fed back to the intelligent control unit. The intelligent control unit optimizes the control strategy by combining the prediction results of the neural network and historical data. The control instructions are transmitted to each device through the execution module to achieve real-time adjustment and dynamic optimization. Through the collaborative work of the above modules, the system can efficiently capture CO2 in ship exhaust while minimizing energy consumption, achieving the dual goals of environmental protection and energy saving.

[0144] This implementation also provides the following experimental data and analysis:

[0145] 1. Cooling process optimization experiment

[0146] Experimental purpose: To verify the effect of temperature control in the cooling process and the accuracy of neural network predictions constrained by physical laws.

[0147] Experimental setup: Changing the exhaust gas flow rate (50-200Nm 3 / h) and initial temperature (150℃-300℃), and record the exhaust gas temperature and energy consumption after cooling.

[0148]

[0149] Table 1 shows the cooling energy consumption and temperature control effect

[0150] Analysis and conclusion: The deviation between the exhaust gas temperature after cooling and the predicted value is controlled within ±0.3℃, indicating that the neural network prediction has high accuracy. Dynamically adjusting the cooling temperature reduces energy consumption by an average of 17.65%, especially under high flow rate and high initial temperature conditions.

[0151] 2. Verification of adsorption efficiency

[0152] Experimental purpose: To test the adsorption efficiency of amine-modified zeolites and metal-organic frameworks (MOFs) and the accuracy of neural network prediction of adsorption efficiency.

[0153] Experimental setup: At different exhaust gas flow rates (50-200Nm 3 / h) and temperature (40-70°C), the adsorption efficiency was measured and compared with the value predicted by the neural network.

[0154]

[0155] Table 2 is a comparison table of adsorption efficiency improvement

[0156] Analysis and conclusion: The neural network has a high prediction accuracy for adsorption efficiency, with a deviation of 0.2%-0.3%. After using the new adsorption material, the adsorption efficiency increased by an average of 9.3%, and it still maintained high performance at high flow rates. The improvement in adsorption efficiency significantly reduced the need for recycling of uncaptured exhaust gas and saved energy.

[0157] 3. Verification of reduced desorption energy consumption

[0158] Experimental purpose: To verify the effect of vacuum desorption method and temperature difference switching method on reducing desorption energy consumption, as well as the accuracy of predicting desorption pressure.

[0159] Experimental setup: Desorption pressure and energy consumption were recorded at different desorption temperatures (100-150 °C) and adsorbent conditions.

[0160]

[0161] Table 3 is a comparison table of desorption energy consumption.

[0162] Analysis and conclusion: The vacuum desorption method and the temperature difference switching method significantly reduce the energy consumption of the desorption process, with an average reduction of 21.25%. The desorption pressure prediction deviation is less than 0.3 kPa, indicating that the intelligent prediction has high reliability. Optimizing the desorption conditions not only reduces energy consumption, but also improves the regeneration efficiency of the desorption material.

[0163] 4. Intelligent control optimization verification

[0164] Purpose of the experiment: To test the effectiveness of the intelligent control unit in dynamically optimizing system parameters, including energy consumption, equipment life, and operational stability.

[0165] Experimental setup: The system energy consumption and failure rate under non-optimization control and neural network optimization control are recorded respectively.

[0166] System control mode Average energy consumption (MJ / h) Failure rate (%) Adsorption material replacement cycle (days) Total energy consumption reduction (%) No optimization control 45.6 2.5 50 Neural Network Optimization Control 38.2 1.1 70 16.2

[0167] Table 4 is a comparison table of energy consumption and failure rate.

[0168] Analysis and conclusion: The intelligent control unit reduces the system energy consumption by 16.2% through dynamic tuning and prolongs the replacement cycle of the adsorption material. The failure rate is reduced by 56%, and the system operation is more stable. The dynamic optimization strategy using neural networks significantly improves the overall performance of the system.

[0169] The experimental conclusion is that the optimization of cooling, adsorption and desorption has significantly reduced energy consumption and improved system efficiency. The prediction model of neural network combined with physical law constraints has high accuracy and provides a reliable basis for dynamic tuning. The energy consumption of the entire system is reduced by about 20% compared with the traditional solution, the adsorption efficiency is increased to more than 90%, and the failure rate is reduced by more than 50%. This technical solution is suitable for actual ship exhaust treatment conditions, meets the goal of high efficiency and energy saving, and has good application prospects.

[0170] Although the present invention has been disclosed as above with preferred embodiments, it is not intended to limit the present invention. A person with ordinary knowledge in the technical field to which the present invention belongs may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be determined by the definition of the claims.

Claims

1. A highly efficient and energy-saving ship exhaust carbon capture system, characterized by: include: An exhaust gas pretreatment module, the exhaust gas pretreatment module comprising a high temperature filter and a condensing device, the high temperature filter is used to remove particulate matter and impurities in the exhaust gas, and the condensing device cools the exhaust gas to a temperature range suitable for treatment through a heat exchanger with an adjustable condensing temperature; A carbon capture module, the carbon capture module comprising an adsorption tower and an adsorption layer plate, the adsorption layer plate being arranged in the adsorption tower, and the adsorption layer plate being provided with an adsorption material for capturing carbon dioxide; A carbon separation and recovery module, comprising a desorption tower and a carbon dioxide compression unit. The desorption tower releases CO2 from the adsorption material using a low-energy vacuum desorption method or a temperature difference switching method, and the carbon dioxide compression unit compresses the captured CO2 into a high-density liquid state; An intelligent control unit includes a central processing unit, a data acquisition module, a communication module, an execution module and a data storage module. The central processing unit is an industrial-grade embedded processor. The central processing unit is used to run a neural network model algorithm to process data in real time and output control instructions. The data acquisition module collects sensor data. The communication module is responsible for real-time communication between sensors, execution devices and the central processing unit. The execution module transmits control instructions to the execution device, which is composed of a multi-channel relay or electric valve control system.

2. The highly efficient and energy-saving ship exhaust carbon capture system according to claim 1, characterized in that: The adsorption material is amine-modified zeolite or metal organic framework material, MOFs.

3. The highly efficient and energy-saving ship exhaust carbon capture system according to claim 1, characterized in that: The communication module is one of CAN bus, Modbus or Ethernet.

4. The highly efficient and energy-saving ship exhaust carbon capture system according to claim 3 is characterized by: The intelligent control unit runs a multi-task neural network with shared features, and the input vector x of the multi-task neural network with shared features includes the key features of all steps: x=[ΔP,PM,V,T in ,T target ,Q,W,C in ,T abs ,P abs ,R regen ,...]; The shared feature multi-task neural network includes a shared feature extraction layer: h=f shared (W shared ·x+b shared ), Wshared , b shared are the weight and bias of the shared network respectively, h: global feature vector; The multi-task neural network with shared features performs the following tasks, where each task builds a branch network based on the shared features: Exhaust cooling temperature prediction task; CO2 adsorption efficiency prediction task; Desorption pressure prediction task; System timing prediction task.

5. The highly efficient and energy-saving ship exhaust carbon capture system according to claim 1, characterized in that: The neural network formula for the exhaust gas cooling temperature prediction task is: T out =f cooling (W cooling ·h+b cooling ); T out : Exhaust gas temperature after cooling, W cooling and b cooling are the weights and bias parameters of the cooling task network, respectively, h: shared input features, including the initial exhaust temperature T in , flow rate V, heat transfer efficiency parameters; Constrained by physical laws, the cooling process follows the heat exchange equation: Q=m·c p ·(T in -T out )Q: heat transfer, m: exhaust gas mass flow, C p : Specific heat capacity of exhaust gas, T in and T out are the initial and final temperatures of the exhaust gas, respectively; Add constraints to the loss function: L cooling =MSE(T out , T out，trute )+λ1·|QQ true |; MSE: Mean square error of cooling temperature prediction, Q true : The actual heat transfer calculated from the measured value, λ1: Physical constraint weight.

6. The highly efficient and energy-saving ship exhaust carbon capture system according to claim 5, characterized in that: The neural network formula for the CO2 adsorption efficiency prediction task is: Adsorption efficiency, W adsorption and b adsorption are the weight and bias parameters of the adsorption efficiency prediction network, h: shared features, including the tower temperature T abs , flow rate Q, adsorption material state R adsorb ; Constrained by physical laws, the adsorption process follows the Lagrangian adsorption isotherm: The partial pressure of CO2 in the tail gas, k: equilibrium constant of the adsorption material; Add constraints to the loss function: MSE: mean square error of adsorption efficiency prediction, λ2: physical constraint weight.

7. The high-efficiency and energy-saving ship exhaust carbon capture system according to claim 6, characterized in that: The neural network formula for the desorption pressure prediction task is: P regen =f desorption (W desorption ·h+b desorption ); P regen : desorption pressure, W desorption and b desorption are the weight and bias parameters of the desorption pressure prediction network, respectively, h: shared features, including the desorption temperature T regen , adsorbent state R adsorb ; Constraints of physical laws: The desorption process conforms to the gas state equation: P regen V = n R T regen ; V: fixed volume of the desorption process, n: number of CO2 molecules adsorbed, R: ideal gas constant, T regen : Desorption temperature, add constraints to the loss function: L desorption =MSE(P regen ,P regen，true )+λ3·|P regen .V-n·R·T regen |; MSE: mean square error of desorption pressure prediction, λ3: physical constraint weight.

8. The high-efficiency and energy-saving ship exhaust carbon capture system according to claim 7, characterized in that: The neural network formula for the system timing prediction task is: t =LSTM(x t ,h t-1 , c t-1 ); h t : The state vector of the current time step, x t : Input characteristics of time step t, including flow rate V, temperature T, equipment state parameters, h t-1 and c t-1 are the hidden layer and cell states of the previous time step respectively; Physical law constraints: system energy consumption E and flow rate V, temperature T and adsorption efficiency There is a relationship: 77: System energy conversion efficiency; Add constraints to the loss function: λ4: physical constraint weight; The final total loss function is the weighted sum of multiple subtasks, with dynamic physical law constraints added: L i : The loss function for each task, α i : Dynamic weight, adjusted with system performance and priority. The physical law constraint term is an important part of the loss function of each subtask, which is used to improve the physical consistency and prediction accuracy of the model.

9. The highly efficient and energy-saving ship exhaust carbon capture system according to claim 8, characterized in that: The data acquisition module comprises: A first temperature sensor, which is installed before the tail gas enters the condensing device and at the outlet of the condensing device. The first temperature sensor monitors the initial temperature and the temperature after cooling of the tail gas to ensure that the tail gas is in a suitable temperature range when entering the adsorption tower; A first gas flow rate sensor, which is installed before the exhaust gas enters the condensing device and at the outlet of the condensing device, and is used to measure the exhaust gas flow rate; A particle sensor is installed after the high-temperature filter to detect the particle content in the exhaust gas and evaluate the filtering effect; The second temperature sensor is installed at each layer of the adsorption layer plate inside the adsorption tower to monitor the temperature distribution inside the adsorption tower to ensure that the adsorption material operates within the optimal working temperature range; A second gas flow rate sensor, which is installed at the inlet and outlet of the adsorption tower to measure the tail gas flow rate and provide a reference for predicting the adsorption efficiency; A first gas composition sensor, which is installed at the entrance of the adsorption tower, between each adsorption layer plate, and at the exit, and is used to monitor the change of CO2 concentration and calculate the adsorption efficiency; A third temperature sensor is installed inside and at the outlet of the desorption tower to monitor the temperature change during the desorption process and optimize the desorption conditions; A pressure sensor is installed inside the desorption tower, at the outlet, and before the compression unit to monitor the desorption pressure and guide the system to optimize operation; A second gas composition sensor, which is installed at the outlet of the analysis tower and is used to monitor the purity of CO2 to ensure the quality of the output gas; Energy consumption sensors are installed at the power interfaces of the cooling device, adsorption tower, analytical tower and compression unit to record the energy consumption of each module in real time and optimize the system operation efficiency.

10. The high-efficiency and energy-saving ship exhaust carbon capture system according to claim 9, characterized in that: The execution module includes an electric valve, a multi-channel relay, a regulating heat exchanger and a pressure regulator. The electric valve controls the exhaust gas flow and direction. The electric valve is arranged before and after the condensing device, at the inlet and outlet of the adsorption tower, and at the outlet of the analysis tower. The multi-channel relay controls the switches of the cooling device, the heater and the compression unit. The multi-channel relay is installed at the power interface of each device. The regulating heat exchanger adjusts the cooling efficiency of the condensing device and is installed inside the condensing device. The force regulator controls the pressure change in the analysis tower and is installed inside the analysis tower.

Images