Method and device for dynamically purifying tail gas of ship

Abstract

The invention discloses a ship tail gas dynamic purification method and device. A real-time monitoring network is constructed by collecting multi-point sensing data of ship tail gas and operation parameters of a purification device, a purification bottleneck area is recognized based on a closed-loop control loop, and encrypted arrangement of sensors is implemented to generate an encrypted monitoring field; collecting monitoring data from the encrypted monitoring field, performing disturbance compensation to form a steady-state measurement value, and constructing an adaptive control domain through the steady-state measurement value; absorbent activity evaluation is carried out on the self-adaptive control domain to extract attenuation characteristics, and the attenuation characteristics are reversely mapped into supplementary dosage to generate a feedforward control signal; analyzing and identifying a speed control step through microcell reaction kinetics, applying temperature regulation and control to the speed control step to generate activation energy compensation, and converting the activation energy compensation into a heating power instruction to form a thermal management sequence; a cooperative control sequence is generated based on coupling of a thermal management sequence and a stoichiometric balance table, the cooperative control sequence is decomposed into a fast and slow response control loop, and efficient and stable purification of ship exhaust is achieved.

Description

Technical Field

[0001] This invention relates to the field of environmental pollution control technology, and in particular to a method and apparatus for dynamic purification of ship exhaust gas. Background Technology

[0002] Ship exhaust gas treatment under complex marine conditions has always been a challenge in the field of environmental protection technology. Ships face multiple disturbances during navigation, including wave rolling, load fluctuations, and environmental changes, posing a severe challenge to the adaptability and stability of exhaust gas purification devices. Traditional purification equipment performs well under fixed land-based conditions, but often experiences performance degradation and control failures in the dynamic marine environment, making it difficult to guarantee continuous and stable purification effects.

[0003] Existing ship exhaust gas purification devices generally employ sparse sensor layouts and coarse-grained parameter control modes, lacking a deep understanding of the internal mechanisms of the purification process. In terms of mass transfer efficiency assessment, traditional methods primarily rely on the concentration difference between inlet and outlet to determine purification effectiveness, failing to accurately identify weak points in the equipment's performance. Regarding absorbent management, existing systems mostly employ timed and quantitative dosing strategies, making it difficult to accurately predict demand based on activity loss. In terms of reaction condition control, traditional temperature regulation mainly relies on empirical parameters, lacking targeted optimization based on local reaction mechanisms. Therefore, a method is urgently needed to address at least one of the above-mentioned problems. Summary of the Invention

[0004] This invention discloses a method and apparatus for dynamic purification of ship exhaust gas. It aims to accurately identify the bottleneck area of ​​purification by constructing a high-density encrypted monitoring field, deeply analyze the limiting links in the purification process by mass transfer resistance analysis and rapid control step positioning technology, achieve multi-parameter collaborative optimization by combining activation energy compensation and stoichiometric balance control, and ensure the stable and efficient operation of the system under complex sea conditions by cascade adjustment of fast and slow response control loops.

[0005] The first aspect of this invention provides a method for dynamic purification of ship exhaust gas, comprising the following steps: Collect multi-point sensor data of ship exhaust gas and operating parameters of purification device. The multi-point sensor data includes pollutant concentration distribution and temperature field distribution. Extract time-domain features from the multi-point sensor data to construct a real-time monitoring network. Establish a closed-loop control loop based on the real-time monitoring network and the operating parameters of the purification device. Mass transfer resistance analysis is performed based on the closed-loop control loop to identify the purification bottleneck area. Sensors are densified in the purification bottleneck area to generate a densified monitoring field. Monitoring data is collected from the densified monitoring field and disturbance compensation is performed to form steady-state measurement values. An adaptive control domain is constructed using the steady-state measurement values. The absorbent activity of the adaptive control domain is evaluated to extract attenuation characteristics. The attenuation characteristics are then mapped back to the supplementary dosage to generate a feedforward control signal. The feedforward control signal is used to trigger precise dosing to form a concentration matching curve. A stoichiometric balance table is then established based on the concentration matching curve. Micro-region reaction kinetics analysis is performed through the encrypted monitoring field to identify rate control steps. Temperature regulation is applied to the rate control steps to generate activation energy compensation. The activation energy compensation is converted into heating power commands to form a thermal management sequence. Based on the thermal management sequence and the stoichiometric balance table, a collaborative control sequence is generated. The coordinated control sequence is decomposed into a fast-response control loop and a slow-response control loop. PID parameters are configured for the fast-response control loop to generate a transient adjustment signal. The transient adjustment signal drives the slow-response control loop to form a cascade adjustment sequence. The cascade adjustment sequence is then used to generate the final purification control command.

[0006] A second aspect of the present invention provides a dynamic purification device for ship exhaust gas, comprising: The data acquisition module is used to collect multi-point sensor data of ship exhaust gas and operating parameters of the purification device. The multi-point sensor data includes pollutant concentration distribution and temperature field distribution. The module extracts time-domain features from the multi-point sensor data to construct a real-time monitoring network. Based on the real-time monitoring network and the operating parameters of the purification device, a closed-loop control loop is established. The bottleneck identification module is used to identify the purification bottleneck area by mass transfer resistance analysis based on the closed-loop control loop, implement a dense sensor arrangement in the purification bottleneck area to generate a dense monitoring field, collect monitoring data from the dense monitoring field, perform disturbance compensation to form a steady-state measurement value, and construct an adaptive control domain through the steady-state measurement value. The feedforward generation module is used to evaluate the absorbent activity of the adaptive control domain, extract the attenuation characteristics, back-map the attenuation characteristics to generate a feedforward control signal for supplementary dosage, use the feedforward control signal to trigger precise dosing to form a concentration matching curve, and establish a stoichiometric balance table based on the concentration matching curve. The temperature control compensation module is used to identify the rapid control step through micro-region reaction kinetic analysis in the encrypted monitoring field, apply temperature regulation to the rapid control step to generate activation energy compensation, convert the activation energy compensation into heating power command to form a thermal management sequence, and couple the thermal management sequence with the stoichiometric balance table to generate a collaborative control sequence. The cascade control module is used to decompose the cooperative control sequence into a fast-response control loop and a slow-response control loop, configure PID parameters for the fast-response control loop to generate a transient control signal, drive the slow-response control loop to form a cascade control sequence through the transient control signal, and use the cascade control sequence to generate the final purification control command.

[0007] The beneficial effects of this invention are reflected in the following points: 1. By using multi-point sensor data acquisition and encrypted monitoring field technology, the dead zone and reaction stagnation zone inside the purification device are accurately located, solving the problem that traditional coarse monitoring cannot detect local performance defects. This enables the system to specifically identify and address weak links in the purification process, improving the monitoring accuracy and control precision under complex conditions such as ship swaying. 2. By employing attenuation characteristic reverse mapping and activation energy compensation technology, precise matching control of absorbent dosage and reaction temperature is achieved. The timing of addition is determined through loss gradient analysis, and reaction conditions are optimized through barrier height regulation, solving the problems of blind addition and temperature regulation lag in traditional methods, ensuring that the purification reaction is always maintained at a high efficiency. 3. Through the cascade adjustment mechanism of fast-response control loop and slow-response control loop, layered processing of disturbances at different time scales is achieved. The fast-response loop handles second-level fluctuations, and the slow-response loop handles minute-level changes. Combined with transient adjustment signals and reverse compensation technology, the impact of ship swaying on the purification effect is effectively suppressed, solving the technical problems of response lag and weak anti-interference ability of traditional single-loop control methods, and improving the operational stability of the ship exhaust gas purification system.

[0008] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0009] The accompanying drawings illustrate specific examples of the technical solutions described in this invention and, together with the detailed embodiments, form part of the specification, serving to explain the technical solutions, principles, and effects of this invention.

[0010] Unless otherwise specified, the same reference numerals in different figures represent the same or similar technical features, and different reference numerals may be used to represent the same or similar technical features.

[0011] Figure 1 This is a schematic flowchart of a dynamic purification method for ship exhaust gas according to the present invention.

[0012] Figure 2 This is a structural block diagram of a dynamic purification device for ship exhaust gas according to the present invention. Detailed Implementation

[0013] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0014] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0015] The technical solutions of the embodiments of this application will be described below.

[0016] like Figure 1 As shown, this embodiment of the invention provides a method for dynamic purification of ship exhaust gas, including the following steps S110-S150: Step S110: Collect multi-point sensor data of ship exhaust gas and operating parameters of purification device. The multi-point sensor data includes pollutant concentration distribution and temperature field distribution. Extract time-domain features from the multi-point sensor data to construct a real-time monitoring network. Establish a closed-loop control loop based on the real-time monitoring network and the operating parameters of purification device.

[0017] Specifically, multi-point sensor data and purification device operating parameters of ship exhaust gas are collected. Sensor arrays are deployed at different cross-sectional locations in the ship's exhaust gas pipeline to continuously monitor the distribution of pollutant concentrations and temperature fields in the exhaust gas. Pollutant concentration monitoring utilizes electrochemical sensors and an infrared absorption spectrometer. The electrochemical sensors detect changes in the concentrations of sulfur dioxide, nitrogen oxides, and carbon monoxide, while the infrared spectrometer measures the distribution of particulate matter and organic pollutants. The sensors are installed in a grid array across the pipeline cross-section, forming a multi-point sampling network. Temperature field distribution monitoring employs thermocouple temperature sensors and an infrared thermal imager. Thermocouple sensors measure the temperature at specific points within the pipeline, while the infrared thermal imager acquires an image of the temperature distribution across the entire cross-section. The data acquisition frequency is set to a high-frequency continuous mode to ensure the capture of dynamic changes in exhaust gas composition and temperature. Simultaneously, purification device operating parameters are collected, including the scrubbing tower's spray water flow rate, spray pressure, circulating pump power, absorbent concentration, and pH value. The temperature, pressure, catalyst activity indicators, and reaction gas flow rate of the catalytic reactor are also included in the monitoring scope. The bag filter dust collector's filtration differential pressure, cleaning cycle, filter bag temperature, and dust collection volume are recorded in real time using dedicated sensors. All sensor data is transmitted to a data acquisition system via fieldbus. The acquisition system is equipped with a high-precision analog-to-digital converter and a data synchronization module to ensure the consistency of timestamps and measurement accuracy of multi-source data.

[0018] A real-time monitoring network is constructed by extracting temporal features from multi-point sensor data. Temporal analysis is performed on pollutant concentration distribution data to calculate statistical characteristics such as the time mean, variance, peak value, and rate of change of concentration. A sliding window technique is used to extract short-term features of concentration changes, with the window length determined based on the time scale of pollutant diffusion. Heat transfer features are extracted from temperature field distribution data, calculating temperature gradient, heat flux density, and temperature distribution uniformity indices. Frequency domain analysis methods are used to identify periodic variation patterns in the multi-point sensor data, and the main frequency components are extracted using Fast Fourier Transform. The data correlation between different sensor locations is analyzed to identify the spatial propagation patterns of pollutant concentration and temperature field. A temporal feature vector F(t) = [C_mean(t), C_var(t), T_grad(t), Q_flow(t)] is constructed, where C_mean is the concentration mean, C_var is the concentration variance, T_grad is the temperature gradient, and Q_flow is the heat flux. The extracted temporal features are used as network nodes, and network connections are determined based on the temporal correlation and spatial coupling relationships between features. A graph theory approach is used to construct a real-time monitoring network topology, where network nodes represent monitoring points, network edges represent data transmission paths, and edge weights reflect the importance of information transmission. The operational status of the entire exhaust gas system is monitored by the state changes of network nodes.

[0019] A closed-loop control loop is established based on a real-time monitoring network and the operating parameters of the purification device. The constructed real-time monitoring network serves as the feedback signal source, forming a closed-loop control loop with the purification device's operating parameters. A proportional-integral-derivative (PI-DE) control strategy is employed, calculating control commands based on pollutant concentration and temperature deviations fed back from the monitoring network. The closed-loop control loop aims to minimize pollutant emission concentration and energy consumption. When the real-time monitoring network detects that the pollutant concentration exceeds a set threshold, the closed-loop control loop automatically adjusts the scrubbing tower's spray flow rate and absorbent concentration. When an abnormal temperature field distribution occurs, the closed-loop control loop adjusts the reaction temperature and gas flow rate of the catalytic reactor. The control parameters of the closed-loop control loop are determined using root locus method and frequency domain analysis to ensure stable operation under various conditions. Abnormal propagation paths are identified through network edge weights to quickly locate the source of the fault. When equipment malfunction or exceeding safe limits is detected, the closed-loop control loop activates a protection mechanism to stop the operation of the relevant equipment. The closed-loop control circuit outputs control commands in a standardized format, including information such as equipment identification, parameter adjustment amount, execution time and priority, to achieve precise automatic control of the purification device.

[0020] Step S120: Based on the mass transfer resistance analysis of the closed-loop control loop, identify the purification bottleneck area, implement a dense sensor arrangement in the purification bottleneck area to generate a dense monitoring field, collect monitoring data from the dense monitoring field, perform disturbance compensation to form steady-state measurement values, and construct an adaptive control domain through the steady-state measurement values.

[0021] Mass transfer resistance analysis based on a closed-loop control loop is used to identify purification bottlenecks. The feedback information provided by the closed-loop control loop includes the operating status and performance parameters of each purification unit; analysis of these parameters identifies areas of abnormal purification efficiency. Pollutant removal rate data from the scrubbing tower is collected, and the inlet and outlet concentration differences are compared with design specifications. Temperature distribution and reaction efficiency in the catalytic reactor are monitored to identify areas of incomplete reaction and abnormal temperatures. The filtration effect of the bag filter is checked, and the change in particulate matter concentration before and after the filter bag is measured. Energy consumption data for each purification unit is recorded, and energy consumption per unit throughput is calculated. Current operating parameters are compared with historical best operating conditions to identify areas where performance degradation exceeds 20%. Spatial analysis methods are used to identify the geographical distribution characteristics of performance-abnormal areas, track the temporal changes in equipment operating status, and observe the development trend of performance degradation. Areas with purification efficiency 15% lower than the design value are designated as candidate bottleneck areas, while also considering areas where energy consumption increases by more than 30%. The spatial extent and impact of each candidate area are statistically analyzed, and the proportion of each area's impact on the overall purification effect is calculated. Cluster analysis is used to group performance-abnormal spatial areas into purification bottleneck areas, with each bottleneck area exhibiting similar performance characteristics and spatial continuity. Prioritize bottleneck areas based on their impact on performance, addressing areas with the greatest impact first. Analyze the main causes of bottlenecks, including equipment wear, deviations in process parameters, and changes in environmental conditions. Record the operating time and maintenance history of equipment within the bottleneck area to identify the degree of equipment aging. The identification results of the purification bottleneck area include the spatial coordinate range of the bottleneck area, its performance level, and the proportion of purification efficiency loss it represents.

[0022] In some embodiments, the step of generating an encrypted monitoring field by encrypting the sensors in the purification bottleneck area includes: generating a swing correction distribution map using the purification bottleneck area; forming a monitoring point-correction correspondence based on the swing correction distribution map; selecting high-gradient monitoring areas from the monitoring point-correction correspondence; and forming an encrypted monitoring field through the encrypted deployment of the high-gradient monitoring areas.

[0023] A roll correction distribution map is generated using the purification bottleneck area. The flow field and mass transfer processes within the purification bottleneck area are affected by the ship's roll motion at sea, causing periodic fluctuations and offsets in the measurement data. The frequency and amplitude characteristics of the ship's roll motion are analyzed, including motion parameters in the three directions of roll, pitch, and heave. The degree of influence of the roll motion at different locations within the purification bottleneck area is calculated, with the impact being more significant at locations farther from the ship's center of gravity. Based on the geometry and device layout of the purification bottleneck area, the distribution of inertial and centrifugal forces caused by the roll motion is calculated. A calculation model for the roll correction coefficient is established: C(x,y,z,t)=A·sin(ωt+φ)+B, where C is the correction coefficient, A is the roll amplitude, ω is the roll frequency, φ is the phase angle, and B is the offset. The purification bottleneck area is spatially gridded, and the roll correction coefficient at each grid point is calculated, forming a three-dimensional correction distribution. The roll correction distribution map displays the spatial variation of the correction coefficient using isosurfaces and color coding, with red areas representing locations with larger correction coefficients and blue areas representing locations with smaller correction coefficients. The swing correction distribution map provides an important reference for subsequent monitoring point layout and data correction.

[0024] A monitoring point-correction correspondence is established based on the swing correction distribution map. Based on the generated swing correction distribution map, corresponding correction parameters and methods are determined for each monitoring point in the densified monitoring field. Correction coefficient values ​​for each candidate monitoring point location are extracted from the swing correction distribution map, serving as an indicator of the swing influence intensity at that point. Three-dimensional interpolation is performed on the swing correction distribution map based on the spatial coordinates of the monitoring points to obtain accurate correction coefficient values. Considering the measurement type and sensor characteristics of the monitoring points, corresponding correction weights are assigned to different types of sensors. Concentration sensors are mainly affected by flow field swing, temperature sensors are mainly affected by changes in thermal conductivity, and pressure sensors are mainly affected by changes in static pressure. Based on the gradient information of the swing correction distribution map, the rate of change of correction coefficients around the monitoring points is calculated as a stability reference for data processing. Monitoring points with similar correction coefficients are grouped, and monitoring points within the same group are treated with the same correction strategy and parameters. The monitoring point-correction correspondence is stored in the form of a data table, containing information such as monitoring point number, spatial coordinates, correction coefficient, correction type, and processing priority. Through this correspondence, the data from each monitoring point can be accurately corrected for swing influence.

[0025] High-gradient monitoring areas are selected from the monitoring point-calibration correspondence. The difference and gradient of the calibration coefficients between adjacent monitoring points are calculated; areas with large gradients indicate drastic changes in the swaying effect, requiring denser monitoring. A gradient threshold screening method is used, setting a threshold for the calibration coefficient gradient; areas exceeding the threshold are marked as high-gradient monitoring areas. High-gradient monitoring areas are typically located at the boundaries of purification bottlenecks, areas with drastic geometric changes, and locations of abrupt changes in flow field characteristics. Cluster analysis is performed on high-gradient monitoring areas, grouping spatially adjacent high-gradient points into contiguous monitoring regions. The area size and gradient intensity of each high-gradient monitoring area are calculated as important parameters for sensor densification. High-gradient monitoring areas are prioritized based on their importance and impact, with higher-priority areas receiving priority for sensor densification. The selection results for high-gradient monitoring areas include the spatial extent of the area, gradient intensity level, monitoring priority, and recommended sensor density. These high-gradient monitoring areas are key areas for sensor densification, significantly improving monitoring accuracy and data reliability.

[0026] For example, the process of forming an encrypted monitoring field through the encrypted deployment of the high gradient monitoring area includes: identifying flow field dead zones based on the high gradient monitoring area; setting activation nodes at the boundaries of the flow field dead zones; collecting dead zone revival signals from the activation nodes to form an effective monitoring range; and adjusting the generation of the encrypted monitoring field through the complementary relationship of the effective monitoring range.

[0027] Dead zones in the flow field are identified based on high-gradient monitoring zones. Due to complex geometry and uneven flow field distribution, dead zones with extremely low or near-stagnant flow velocities exist within high-gradient monitoring zones. By calculating the velocity vector and streamline distribution at various points within the high-gradient monitoring zone, areas with flow velocities below a set threshold are identified. The criteria for identifying dead zones are flow velocities less than 10% of the average velocity and frequent or no clear flow direction. The causes of dead zone formation are analyzed, including flow separation caused by geometric structures, backflow caused by pressure gradients, and boundary layer separation caused by viscous effects. The geometric parameters of the dead zones are calculated, including their spatial extent, volume, and shape characteristics. Mass transfer efficiency is extremely low within dead zones, and pollutants easily accumulate, making them a weak point in purification. Dead zones are classified according to their causes and characteristics into separation-type dead zones, backflow-type dead zones, and stagnant dead zones. The identification results of dead zones include their spatial coordinates, type, degree of impact, and improvement suggestions. These dead zones in the flow field are key areas for intensive monitoring and require effective monitoring through special monitoring methods.

[0028] Activation nodes are set at the boundaries of the dead zones in the flow field. These nodes enhance the monitoring capability of the dead zones. The boundaries of the dead zones are the intersection of the dead zone and the active flow field, a critical area for pollutant exchange and mass transfer. The activation nodes employ highly sensitive sensors capable of detecting subtle concentration changes and flow signals. Inflow activation nodes are set at the upstream boundary of the dead zone to monitor the flow rate and concentration of pollutants entering the dead zone. Outflow activation nodes are set at the downstream boundary of the dead zone to monitor the pollutants flowing out of the dead zone and their residence time. Considering the three-dimensional geometry of the dead zones, activation nodes are also set at the upper and lower boundaries, forming a three-dimensional, encircling monitoring network. The sensors of the activation nodes are miniaturized to reduce interference and impact on the flow field. The activation nodes are connected via a communication network to achieve real-time data sharing and collaborative monitoring. The data acquisition frequency of the activation nodes is higher than that of conventional sensors, enabling them to capture transient changes at the boundaries of the dead zones. Through the collaborative work of the activation nodes, effective encirclement and comprehensive monitoring of the dead zones in the flow field are achieved.

[0029] The effective monitoring range is formed by acquiring dead zone reactivation signals from active nodes. Monitoring signals from the dead zone of the flow field are acquired from the designated active nodes to identify active signals and reactivation characteristics within the dead zone. Dead zone reactivation signals refer to signals indicating flow recovery, concentration changes, and active mass transfer within the dead zone. Active nodes monitor pressure fluctuations, concentration pulsations, and temperature changes at the dead zone boundary through high-frequency data acquisition. The spectral characteristics of the active node signals are analyzed to identify characteristic frequencies and signal patterns related to dead zone reactivation. When periodic concentration changes or flow signals are detected within the dead zone, it is identified as a dead zone reactivation signal. The intensity and frequency of the dead zone reactivation signal reflect the activity level and mass transfer capacity of the dead zone. Based on the intensity of the dead zone reactivation signal acquired by the active nodes, the effective monitoring spatial range is determined. The effective monitoring range refers to the spatial area where dead zone reactivation signals can be reliably detected; signals outside this range are too weak to be effectively monitored. The boundary line of the effective monitoring range is plotted by calculating the spatial distribution of the signal intensity. The size and shape of the effective monitoring range depend on the propagation characteristics of the dead zone reactivation signal and the detection capability of the active nodes.

[0030] A dense monitoring field is generated by adjusting the complementary relationships of effective monitoring ranges. The overlapping and blank areas between the effective monitoring ranges of different active nodes are analyzed; overlapping areas represent monitoring redundancy, and blank areas represent monitoring blind spots. Redundant monitoring areas are optimized to reduce unnecessary sensor placement, lowering system complexity and cost. Monitoring blind spots are supplemented by adding sensors or adjusting the positions of active nodes in blank areas to eliminate monitoring dead zones. The optimal monitoring coverage and resource allocation are achieved through the complementary relationships of effective monitoring ranges. A mathematical model of the effective monitoring range is established, calculating the intersection, union, and complement relationships between ranges. Based on the analysis results of complementary relationships, the sensor density and position are adjusted to form an optimized monitoring network topology. The final configuration of the dense monitoring field ensures the integrity of monitoring coverage while avoiding resource waste and redundancy. The dense monitoring field has comprehensive monitoring capabilities for dead zones and high-gradient regions in the flow field, providing a reliable data foundation for precise control of bottleneck areas in air purification.

[0031] In some embodiments, the step of collecting monitoring data from the encrypted monitoring field and performing disturbance compensation to form a steady-state measurement value includes: collecting the encrypted monitoring field data and segmenting it into a mainstream signal and a disturbance signal; extracting phase difference features between the mainstream signal and the disturbance signal; constructing a time delay compensation matrix using the phase difference features; and performing phase correction on the disturbance signal according to the time delay compensation matrix to generate a steady-state measurement value.

[0032] Data from the encrypted monitoring field was collected and segmented into mainstream and disturbance signals. The monitoring data from the encrypted field contains key information about the purification process and interference information from various external disturbances. The mainstream signal reflects the normal operating status of the purification device and the pollutant treatment process, exhibiting relatively stable temporal characteristics and spatial distribution. Disturbance signals originate from external factors such as ship swaying, environmental changes, equipment vibration, and measurement noise, exhibiting randomness and fluctuation. Frequency domain analysis was used to decompose the monitoring data, identifying signals of different frequency components through Fast Fourier Transform. The mainstream signal is mainly concentrated in the low-frequency band, corresponding to the slow changes and steady-state characteristics of the purification process. Disturbance signals are mainly distributed in the mid-to-high frequency band, corresponding to rapidly changing external interference and random fluctuations. A frequency threshold was set to separate the monitoring data into low-frequency mainstream signals and high-frequency disturbance signals. The separated signals were then reconstructed in the time domain to obtain independent time series of the mainstream and disturbance signals. The accuracy of the signal separation was verified to ensure that the superposition of the mainstream and disturbance signals could reconstruct the original monitoring data.

[0033] Phase difference features are extracted between the mainstream signal and the disturbance signal. The instantaneous phases of the mainstream and disturbance signals are calculated, and an analytical representation of the signals is obtained using Hilbert transform. The instantaneous phase difference is defined as Δφ(t) = φ_d(t) - φ_m(t), where φ_d is the phase of the disturbance signal and φ_m is the phase of the mainstream signal. The temporal variation law of the phase difference is analyzed to identify its periodicity and trend. Statistical characteristic parameters of the phase difference, including mean, standard deviation, skewness, and kurtosis, are calculated as quantitative indicators of signal coupling strength. The time delay relationship between the mainstream and disturbance signals is calculated using correlation analysis to determine the lag or lead time of the disturbance signal relative to the mainstream signal. The phase difference features reflect the degree and manner of the disturbance's influence on the mainstream signal, serving as an important basis for disturbance compensation. Phase difference features are extracted from sensor signals at different spatial locations, forming a spatial distribution map of the phase difference features. The spatial distribution of the phase difference features reveals the spatial propagation law and local differences of the disturbance's influence.

[0034] A time delay compensation matrix is ​​constructed using phase difference characteristics. The elements of the time delay compensation matrix correspond to compensation coefficients for different sensor positions and different frequency components. The time delay compensation coefficient τ = Δφ / (2πf) is calculated for each sensor position based on the phase difference characteristics, where Δφ is the phase difference and f is the signal frequency. Compensation coefficients are calculated separately for different frequency components, forming a frequency-position two-dimensional compensation matrix. Rows in the time delay compensation matrix correspond to different sensor positions, columns correspond to different frequency components, and matrix elements are the corresponding compensation coefficients. Considering the spatial correlation between sensors, the compensation coefficients of adjacent sensors are smoothed to avoid discontinuities in the compensation process. The parameters of the compensation matrix are adjusted through an iterative optimization method to minimize the variance of the compensated signal. The time delay compensation matrix has adaptive capabilities, able to adjust the compensation parameters in real time according to changes in phase difference characteristics, achieving accurate correction of disturbed signals. The time delay compensation matrix is ​​sparsified to retain compensation coefficients with significant impact, improving computational efficiency.

[0035] Steady-state measurements are generated by phase correction of the disturbance signal using a time delay compensation matrix. The disturbance signal is convolved with the time delay compensation matrix in the frequency domain to adjust the phase of each frequency component. The phase-corrected disturbance signal is then superimposed with the original mainstream signal to obtain the corrected total signal. An inverse Fourier transform is used to convert the corrected frequency domain signal back to the time domain, obtaining the time series of steady-state measurements. The steady-state measurements eliminate the main disturbance effects, preserving the true information and trends of the purification process. Statistical analysis is performed on the steady-state measurements to calculate signal stability and accuracy indices, quantifying the compensation effect. The variance and noise levels of the signal before and after correction are compared to verify the effectiveness of the time delay compensation matrix. The steady-state measurements provide an accurate and reliable data foundation for subsequent control decisions and system optimization. The steady-state measurements are stored in association with their corresponding time labels and spatial coordinates, forming a complete steady-state monitoring database.

[0036] An adaptive control domain is constructed using steady-state measurements. Based on the obtained steady-state measurements, an adaptively adjustable control domain is built to optimize the purification effect. The steady-state measurements contain precise state information of the purification bottleneck area, including key parameters such as pollutant concentration distribution, temperature field distribution, and mass transfer efficiency. The boundary range of the control domain is divided according to the spatial distribution characteristics of the steady-state measurements, with each control domain corresponding to a specific purification bottleneck area. The mean and gradient distribution of the parameters within the control domain are calculated as important inputs for control strategy formulation. Trend analysis is performed on the steady-state measurements to identify the regularity and periodicity of parameter changes over time. The adjustment requirements of the control domain are calculated based on the deviation between the steady-state measurements and the set target values, including the adjustment direction and magnitude of the control parameters. A combination of fuzzy control and neural network control is used to automatically adjust the parameter settings of the control domain according to changes in the steady-state measurements. The control domain has self-learning capabilities and can adjust the control strategy and parameter boundaries based on historical data of steady-state measurements. The adaptive control domain can respond to changes in the state of the purification bottleneck area and adjust the control strategy in real time to maintain the optimal purification effect. The output of the control domain includes specific control commands and parameter adjustment suggestions, which are applied to the operating parameters of the purification device.

[0037] Step S130: The absorbent activity of the adaptive control domain is evaluated and the attenuation characteristics are extracted. The attenuation characteristics are back-mapped to the supplementary dosage to generate a feedforward control signal. The feedforward control signal is used to trigger precise dosing to form a concentration matching curve. A stoichiometric balance table is established based on the concentration matching curve.

[0038] Specifically, the activity of the absorbent in the adaptive control domain was evaluated to extract attenuation characteristics. The adaptive control domain provides precise state information of the purification bottleneck area, including key parameters such as absorbent concentration distribution, pH changes, and temperature distribution. By analyzing the temporal trend of absorbent concentration in the adaptive control domain, the consumption rate and remaining activity of the absorbent were calculated. The pH value of the alkaline absorbent in the scrubbing tower was continuously monitored; the decrease in pH reflects the degree of reaction between the absorbent and acidic pollutants. The absorbent activity attenuation rate A(t) = (A0 - A(t)) / A0 was calculated, where A0 is the initial activity, A(t) is the activity at time t, and the attenuation rate reflects the degree of absorbent failure. The distribution differences of absorbent activity at different spatial locations were analyzed to identify the areas with the most severe activity attenuation. The concentration changes of the effective components in the absorbent were determined using chemical analysis methods to quantify the chemical mechanism of activity loss. The activity of the catalyst in the catalytic reactor was evaluated, and the catalyst activity attenuation status was reflected by conversion efficiency and selectivity indicators. The temporal characteristics of activity attenuation were extracted, including parameters such as attenuation rate, attenuation mode, and attenuation period.

[0039] In some embodiments, the step of inversely mapping the attenuation feature to generate a feedforward control signal by supplementary dosage includes: identifying the activity loss rate of the attenuation feature to generate a loss gradient; dividing the attenuation feature into multiple supplementary intervals to form a tiered demand spectrum based on the loss gradient; using the tiered demand spectrum to match differentiated dosage rates to generate a segmented supplementary sequence; and configuring the segmented supplementary sequence with trigger thresholds to generate a feedforward control signal.

[0040] The degradation characteristics are analyzed to identify the activity loss rate and generate a loss gradient. The degradation characteristics contain information about the temporal and spatial changes in absorbent activity; loss rate analysis quantifies the rate and extent of activity degradation. The activity loss rate is calculated in the time dimension, reflecting the rate of change of activity over time. The activity loss rate in the spatial dimension is also calculated, reflecting the distribution gradient of activity at different spatial locations. Loss rates are calculated separately for different types of absorbents: alkaline absorbents are mainly affected by the consumption of acidic pollutants, while oxidizing absorbents are mainly affected by the consumption of reducing pollutants. Factors influencing the activity loss rate are analyzed, including pollutant concentration, reaction temperature, contact time, and catalyst activity. The distribution characteristics of the loss rate are calculated using statistical analysis methods, identifying the mean, variance, and extreme values. The loss gradient is defined as the spatial rate of change of the activity loss rate, reflecting the spatial non-uniformity of the loss degree. The magnitude and direction of the loss gradient are calculated to identify the regions with the most severe activity loss and the main direction of loss diffusion. Regions with high values ​​in the loss gradient correspond to key locations requiring priority replenishment.

[0041] A tiered demand spectrum is formed by dividing replenishment intervals into multiple levels based on the loss gradient. The magnitude of the loss gradient reflects the urgency of absorbent replenishment in different regions; regions with high gradients require priority and large-scale replenishment. Tiered thresholds are set based on the numerical range of the loss gradient, dividing the replenishment regions into three levels: high-demand, medium-demand, and low-demand zones. High-demand zones correspond to areas with loss gradients greater than the upper threshold, requiring immediate and large-scale replenishment. Medium-demand zones correspond to areas with loss gradients in the medium range, requiring moderate replenishment to maintain activity balance. Low-demand zones correspond to areas with loss gradients less than the lower threshold, requiring only small-scale replenishment or no replenishment at all. The required replenishment amount and frequency are calculated for each replenishment interval, forming quantified demand parameters. The tiered demand spectrum is organized in tabular form, including interval number, loss gradient range, demand level, replenishment amount, and replenishment frequency. Considering the mutual influence and coupling relationships between different intervals, the replenishment parameters for each interval are adjusted to achieve overall balance. The tiered demand spectrum provides detailed parameter guidance for differentiated application, ensuring the rational allocation and efficient utilization of replenishment resources.

[0042] A segmented replenishment sequence is generated by matching differentiated acceleration rates to a tiered demand spectrum. Based on the formed tiered demand spectrum, corresponding differentiated acceleration rates are matched to generate segmented replenishment sequences with a time-sequential arrangement. The tiered demand spectrum provides replenishment demand levels and quantification parameters for different intervals, which need to be converted into specific replenishment operation sequences. Differentiated acceleration rate configurations are determined according to demand levels: a fast replenishment mode is used in high-demand areas, a conventional replenishment mode in medium-demand areas, and a slow replenishment mode in low-demand areas. The replenishment duration T=Q / R for each interval is calculated, where Q is the replenishment amount and R is the acceleration rate. Considering the capacity limitations and safety constraints of the replenishment equipment, the acceleration rate setting is adjusted to meet equipment operating requirements. The time sequence and parallelism of the replenishment operations are determined based on the spatial distribution of the tiered demand spectrum. For adjacent replenishment intervals, a sequential replenishment method is used to avoid mutual interference. For independent replenishment intervals, a parallel replenishment method is used to improve replenishment efficiency. The segmented replenishment sequence contains detailed information such as the start time, end time, acceleration rate, and replenishment location for each replenishment segment. By optimizing the timing of the segmented supplementation sequence using an algorithm, the total dosing time is minimized while achieving the optimal dosing effect. The segmented supplementation sequence provides the automatic dosing system with precise execution instructions and timing control.

[0043] The segmented replenishment sequence is configured with trigger thresholds to generate feedforward control signals. The segmented replenishment sequence provides a detailed arrangement of the dosing operation, which needs to be converted into executable control signals and trigger conditions. A trigger threshold is set for each dosing segment; when the monitored parameter reaches the threshold, the corresponding dosing operation is automatically initiated. Trigger thresholds include various types such as activity concentration threshold, pH threshold, temperature threshold, and time threshold. Different trigger sensitivities are set according to the urgency of the segmented replenishment sequence; a lower trigger threshold is used in high-demand areas to achieve rapid response. A dosing stop condition threshold is configured; the dosing operation is automatically stopped when the target concentration or dosing limit is reached. The feedforward control signal uses a digital signal format and includes information such as device address, operation code, parameter values, and execution time. The feedforward control signal is transmitted to the dosing control system via a communication protocol to achieve remote automatic control. The feedforward control signal has a priority setting function; the emergency dosing signal has the highest priority and can interrupt regular dosing operations. Redundancy and backup mechanisms are configured for the signal to ensure the reliability of control signal transmission and the continuity of dosing operations. The feedforward control signal achieves complete closed-loop control from attenuation characteristic analysis to automatic dosing execution.

[0044] A concentration matching curve is generated by triggering precise dosing using a feedforward control signal. Based on the generated feedforward control signal, the absorbent dosing system is triggered to perform precise dosing operations, forming a concentration matching curve reflecting the dosing effect. The feedforward control signal drives the automatic dosing device to add absorbent according to predetermined parameters, employing a multi-point distributed dosing method. Dosing nozzles are set at different height levels of the scrubbing tower, and differentiated dosing is performed according to the distribution ratio of the feedforward control signal. Changes in absorbent concentration are monitored in real time during dosing, recording concentration data before, during, and after dosing. The concentration of the effective component of the absorbent is continuously measured using an online analyzer, forming a dynamic curve of concentration change over time. The concentration matching curve reflects the correspondence between the dosage and the concentration response, including the rate of concentration rise, the time to reach equilibrium, and the final stable concentration. The concentration response characteristics under different dosages are analyzed to determine the quantitative relationship between dosing efficiency and dosing effect. Concentration changes at multiple dosing points are monitored synchronously, forming a spatially distributed set of concentration matching curves. Mathematical modeling of the concentration matching curves is performed using curve fitting methods to obtain a functional expression for the concentration response.

[0045] A stoichiometric balance table is established based on the concentration matching curve. Concentration values ​​at key time points are extracted from the concentration matching curve, including the initial concentration before addition, the instantaneous concentration during addition, and the final concentration after reaction equilibrium. Stoichiometric coefficients are determined based on the main absorption reaction equations: the stoichiometric ratio for the sulfur dioxide absorption reaction SO2 + Ca(OH)2 → CaSO3 + H2O is 1:1, and the stoichiometric ratio for the nitrogen oxide absorption reaction 2NO2 + Ca(OH)2 → Ca(NO3)2 + H2O is 2:1. The ratio of the actual addition amount to the theoretical requirement is calculated to obtain the stoichiometric efficiency η = C_actual / C_theoretical, where Q_actual is the actual addition amount and Q_theoretical is the theoretical required addition amount. The consumption and production amounts of each component are determined through material balance calculations; the consumed absorption amount equals the initial concentration minus the remaining concentration. A row-column structure is established for the stoichiometric balance table, with rows corresponding to different chemical reactions and columns corresponding to reactants, products, stoichiometric coefficients, and conversion rates. Enter the quantitative data for each reaction, including reactant consumption concentration, product formation concentration, theoretical stoichiometric ratio, and actual conversion rate. Calculate the effects of side reactions and competing reactions, and correct the stoichiometric relationships of the main reactions. Perform a mass conservation check on the balance sheet to ensure that the mass balance relationships of each component conform to the conservation law. The stoichiometric balance sheet fully records the quantitative chemical relationships of the absorption process, including the stoichiometric data and efficiency parameters of the main reactions.

[0046] Step S140: Micro-region reaction kinetics analysis is performed through encrypted monitoring field to identify rate control steps. Temperature regulation is applied to the rate control steps to generate activation energy compensation. The activation energy compensation is converted into heating power command to form a thermal management sequence. Based on the thermal management sequence and the stoichiometric balance table, a collaborative control sequence is generated.

[0047] In some embodiments, the step of identifying rate control through micro-region reaction kinetic analysis using the encrypted monitoring field includes: converting the encrypted monitoring field data into a micro-region reaction rate field; identifying reaction stagnation regions from the micro-region reaction rate field; injecting boundary perturbations into the reaction stagnation regions to form an activation path; and locating rate limiting elements along the activation path to form a rate control step.

[0048] The encrypted monitoring field data is converted into a micro-region reaction rate field. The encrypted monitoring field data contains high spatial resolution information on concentration distribution, temperature distribution, and mass transfer coefficient distribution. The time derivative of the concentration data is calculated to obtain the concentration change rate at each monitoring point. Based on the reaction stoichiometry, the concentration change rate is converted into the reaction rate, considering the influence of the stoichiometry coefficient. The reaction rate of each component in the multi-component reaction system is calculated separately, forming a multi-dimensional rate vector field. Spatial interpolation methods are used to extend the rate data from discrete monitoring points into a continuous rate field distribution. Cubic spline interpolation is used to process the spatial distribution of the rate data to ensure the smoothness and continuity of the rate field. Considering the temperature dependence of the reaction, the numerical values ​​of the rate field are corrected by incorporating temperature field data. The rate field is normalized to eliminate the influence of orders of magnitude differences between different reactions. The micro-region reaction rate field is stored in the form of a three-dimensional scalar field, containing spatial coordinates and corresponding reaction rate values.

[0049] Identifying reaction stagnation regions from micro-region reaction rate fields. Micro-region reaction rate fields display the spatial variation of reaction activity; regions with extremely low rates correspond to stagnation or near-stagnation. A reaction rate threshold is set; when the reaction rate in a region of the rate field is less than 10% of the average of the entire rate field, it is marked as a reaction stagnation region. Connectivity analysis is used to group spatially continuous low-rate regions into complete stagnation regions. Geometric parameters for each reaction stagnation region are calculated, including volume, shape characteristics, and boundary curvature. The causes of reaction stagnation are analyzed, including factors such as mass transfer limitation, insufficient temperature, catalyst deactivation, and geometrical obstacles. Reaction stagnation regions are classified according to their causes into diffusion-limited, kinetic-limited, and mixed-limited types. A severity index for reaction stagnation regions is calculated; a higher index indicates a more severe stagnation. The identification results of reaction stagnation regions include attributes such as stagnation region number, spatial location, volume, stagnation type, and severity.

[0050] Activation paths are formed by injecting boundary perturbations into the reaction stagnation zone. Artificial perturbations are injected at the identified boundary locations of the stagnation zone to activate the reaction processes within it, forming an activation path connecting the active and stagnant regions. The boundary of the stagnation zone is the interface between the active and stagnant regions; boundary perturbations allow the active reaction to propagate into the stagnation zone. Concentration perturbations are injected at the upstream boundary of the stagnation zone to increase reactant concentration and drive the reaction. Temperature perturbations are injected at the lateral boundary to increase local temperature and enhance reactivity. Flow perturbations are injected at the downstream boundary to enhance mass transfer and eliminate diffusion limitations. Perturbation injection is performed in a pulsed manner, with the perturbation intensity gradually increasing until a significant active response appears in the stagnation zone. Rate changes within the stagnation zone after perturbation injection are monitored to identify the most sensitive perturbation locations and types. The propagation process of the perturbation within the stagnation zone is simulated using particle tracking methods, and the spatial trajectory of the perturbation propagation is plotted. The activation path is defined as the spatial channel through which the perturbation can effectively propagate and activate the reaction; the path extends along the direction of the rate gradient within the stagnation zone.

[0051] Rate-limiting steps are identified along the activation path to form rate-controlling steps. Based on the established activation path, detailed kinetic analysis is performed along the path to pinpoint the key rate-limiting steps affecting the reaction rate. The activation path connects the active and stagnant reaction regions, and the rate-limiting steps along the path determine the overall reaction rate characteristics. Analytical nodes are set at fixed intervals along the activation path, and detailed reaction kinetic parameters are calculated at each node. Multi-step reaction decomposition is performed at each analytical node to identify the rate constants and activation energies of each elementary reaction. The reaction rate differences between different nodes are compared, and the location with the most significant rate abrupt change corresponds to the main rate-limiting step. Rate control analysis methods are used to calculate the degree of control of each elementary reaction on the overall reaction rate, and the elementary reaction with the greatest degree of control is the rate-limiting step. The reaction mechanism of the rate-limiting step is analyzed to determine the physicochemical reasons for the rate limitation, including high activation energy barriers, insufficient concentration, low temperature, and low catalytic activity. Detailed information such as the spatial location, reaction type, cause of rate limitation, and improvement direction of the rate-limiting step is recorded. All rate-limiting steps identified along the activation path are integrated into a complete rate-controlling step sequence.

[0052] In some embodiments, applying temperature control to the rate control step to generate activation energy compensation includes: constructing a reaction barrier distribution using the rate control step; performing energy analysis on the reaction barrier distribution to form a barrier height marker; using the barrier height marker to determine a heating zone; and generating activation energy compensation by increasing the temperature of the heating zone.

[0053] The reaction barrier distribution is constructed using rate-determining steps. These steps provide the type, location, and kinetic parameters of the rate-limiting reaction, which form the basis for constructing the barrier distribution. The corresponding barrier height is calculated based on the activation energy of each rate-determining step. For gas-phase reactions, the barrier height is primarily determined by intermolecular collisions and bond breaking and recombination. For liquid-phase reactions, the effects of solvation and ionic strength must also be considered. For gas-liquid interface reactions, the barrier distribution is significantly influenced by interfacial tension and interphase mass transfer resistance. Transition state theory is used to calculate barrier curves along the reaction path, determining the functional relationship between reaction coordinates and barrier height. The barrier data for each rate-determining step are organized according to spatial location to form a three-dimensional barrier distribution field. Barrier interpolation is used to handle the continuity of the spatial distribution, ensuring a smooth transition between barrier values ​​in adjacent regions. The effect of temperature on barrier height is considered; the barrier height decreases with increasing temperature.

[0054] Energy analysis of the reaction barrier distribution is performed to create barrier height markers. The reaction barrier distribution displays the energy barrier height at various points in space and needs to be classified according to differences in barrier height. A classification threshold is set for barrier height, dividing the barrier distribution into three levels: high barrier region, medium barrier region, and low barrier region. The high barrier region corresponds to areas with barrier heights greater than the average plus two standard deviations; these areas have extremely slow reaction rates and require intensive heating. The medium barrier region corresponds to areas with barrier heights near the average; these areas require moderate heating to improve reaction efficiency. The low barrier region corresponds to areas with barrier heights less than the average minus one standard deviation; these areas have relatively active reactions and only require slight heating. The volume proportion and spatial distribution characteristics of each barrier level region are calculated as an important reference for heating resource allocation. Barrier heights are color-coded: red for high barrier regions, yellow for medium barrier regions, and green for low barrier regions.

[0055] Heating zones are determined using barrier height markers. Based on the generated barrier height markers, specific spatial zones requiring temperature control and heating parameters are identified. The barrier height markers provide information on the energy barrier level and heating priority for different regions. Regions marked with high barrier heights are designated as priority heating zones, with the highest heating temperature. Regions marked with medium barrier heights are designated as regular heating zones, with a medium heating temperature. Regions marked with low barrier heights are designated as lightly heated zones, with the lowest heating temperature. The required heating temperature is calculated based on the specific barrier height values ​​to achieve the desired acceleration of the reaction rate. Considering the geometric constraints and equipment layout of the heating zones, the boundary ranges are adjusted to adapt to actual installation conditions. Adjacent heating zones of the same level are merged to reduce the complexity of heating control.

[0056] For example, the activation energy compensation generated by the temperature increase in the heating zone includes: converting the temperature sequence of the heating zone into a temperature gradient distribution; driving the gas-liquid interface update based on the temperature gradient distribution to generate a mass transfer coefficient increment; extracting a reaction acceleration factor from the mass transfer coefficient increment; and generating activation energy compensation according to the cumulative effect of the reaction acceleration factor.

[0057] The temperature sequence of the heating zone is converted into a temperature gradient distribution. By monitoring the temperature changes at various locations within the heating zone, the spatial differences in temperature distribution are analyzed. Temperature sensors are placed at different heights inside the scrubbing tower and at the inlet and outlet of the catalytic reactor to record the time sequence and spatial variation of temperature rise during heating in real time. The temperature difference between adjacent monitoring points is calculated to identify areas of drastic temperature changes and areas with relatively gentle temperature distribution. For example, when the heater power is increased by 20%, the temperature gradient near the heater can reach 15°C / m, while the temperature gradient in corner areas farther from the heating source is only 3°C / m, forming a significant uneven temperature distribution. Locations with large temperature gradients indicate strong heat transfer, while locations with small temperature gradients may have heat transfer obstacles or heating blind spots. Through spatial analysis of the temperature data, a temperature distribution map of the heating zone is drawn, identifying high-temperature gradient regions and low-temperature gradient regions.

[0058] The increase in mass transfer coefficient is generated by the renewal of the gas-liquid interface driven by temperature gradient distribution. The temperature gradient distribution directly alters the physicochemical properties of the gas-liquid interface region, driving the dynamic renewal process of the interface. High-temperature gradient regions reduce the liquid density near the interface, generating buoyancy-driven upward motion, continuously replenishing the gas-liquid interface with fresh liquid. Simultaneously, the increased temperature reduces interfacial tension and liquid viscosity, making the interface more active and deformable. The density difference caused by the temperature gradient creates natural convection near the interface, accelerating the renewal and replacement process of the aging interface. The increased interface renewal frequency directly reduces the thickness of the mass transfer boundary layer; the previously stagnant interface region is replaced by fresh liquid, restoring the driving force for mass transfer. Analysis of interface renewal theory shows that the interface renewal frequency is proportional to the square root of the temperature gradient. Based on the new physical properties after interface renewal, the mass transfer coefficient is recalculated and compared with the value before renewal. The increase in mass transfer coefficient is defined as the mass transfer coefficient after interface renewal minus the mass transfer coefficient before renewal, quantifying the promoting effect of temperature gradient distribution on the mass transfer process. Differences in temperature gradients at different locations result in a non-uniform spatial distribution of the mass transfer coefficient increase, with larger increases in high-temperature gradient regions.

[0059] The reaction acceleration factor is extracted from the mass transfer coefficient increment. Based on the calculated mass transfer coefficient increment, the impact of mass transfer improvement on the overall reaction rate is analyzed, and a quantitative reaction acceleration factor is extracted. The mass transfer coefficient increment affects the rate of reactant transfer to the reaction interface, thus affecting the overall reaction rate. For mass transfer-controlled reactions, the reaction rate is directly proportional to the mass transfer coefficient. For reaction-controlled processes, mass transfer improvement indirectly accelerates the reaction by increasing the reactant concentration. A mass transfer-reaction coupling model is established to analyze the quantitative impact of changes in the mass transfer coefficient on the reaction rate. The reaction acceleration factor AF = (r_enhanced) / (r_original) is calculated, where r_enhanced is the reaction rate after mass transfer improvement, and r_original is the original reaction rate. The acceleration factor for each reaction is calculated according to the degree of mass transfer-reaction coupling for different reactions. Dimensional analysis is used to verify the rationality of the acceleration factor calculation and ensure the physical meaning of the factor values. The time dependence of the acceleration factor is analyzed to identify the establishment time and duration of the acceleration effect. Considering the spatial non-uniformity of mass transfer improvement, the spatial variation of the local acceleration factor is calculated. The results of the reaction acceleration factor extraction include factor value, applicable reaction, spatial range, and temporal characteristics.

[0060] Activation energy compensation is generated based on the cumulative effect of the reaction acceleration factor. The reaction acceleration factor reflects the promoting effect of temperature increase on the reaction rate by improving the mass transfer process. The simultaneous action of multiple acceleration mechanisms produces a cumulative effect, and the total acceleration effect is equal to the product of the individual effects. The total acceleration factor AF_total = ∏AF_i is calculated, where AF_i is the acceleration factor of the i-th acceleration mechanism. According to the inverse operation of the Arrhenius equation, the acceleration factor is converted into an equivalent activation energy compensation. The activation energy compensation is calculated as ΔEa = -R·T·ln(AF_total), where R is the gas constant and T is the average temperature. Considering the difference in acceleration degree of different reaction steps, the activation energy compensation for each step is calculated separately. The activation energy compensation is integrated over time to calculate the total cumulative compensation effect. The generation result of activation energy compensation includes the compensation value, spatial distribution, temporal evolution, and influencing mechanism. Activation energy compensation effectively accelerates the rate-controlling steps, improving the overall purification efficiency.

[0061] The activation energy compensation is converted into a heating power command to form a thermal management sequence. The activation energy compensation amount determines the required heating energy, and the heating power demand is determined through heat balance calculations. The heating power is calculated as P = m·Cp·ΔT / Δt + Q_loss, where P is the heating power, m is the mass flow rate of the heating medium, Cp is the specific heat capacity, ΔT is the temperature rise, Δt is the heating time, and Q_loss is the heat loss. The power demand for each heating zone is calculated based on its geometry and heat transfer characteristics. Considering the power limitations and response characteristics of the heating equipment, the total power demand is allocated to multiple heating units. The heating start-up time and duration are determined based on the reaction progress of the rapid control step, with emergency heating demands having the highest priority and capable of interrupting the regular heating process. The thermal management sequence includes control parameters such as heating unit number, power setpoint, start-up time, duration, and stop conditions. Energy consumption is optimized for the thermal management sequence to minimize total energy consumption while meeting reaction requirements. The heating timing is optimized using predictive control methods, with heating started earlier to compensate for the system's thermal inertia.

[0062] A coordinated control sequence is generated by coupling the thermal management sequence with the stoichiometric balance sheet. The heating power data from the thermal management sequence is compared and analyzed with the dosage data from the stoichiometric balance sheet to determine the timing matching relationship between the two. When the stoichiometric balance sheet shows a 20% increase in SO2 absorbent dosage, the heating power of the scrubbing tower is simultaneously increased by 15% to maintain the reaction temperature. When the balance sheet shows NO... X When the reaction efficiency decreases by more than 10%, the heating power of the catalytic reactor is increased by 25% to accelerate the reaction. A time interval between addition and heating is set; the corresponding heating operation is initiated 5 minutes after the addition operation to avoid temperature shock. A proportional relationship for power adjustment is established: for every 1 kg / h increase in addition, the corresponding heating power increases by 50 kW. The power allocation ratio of the thermal management sequence is adjusted according to the reaction stoichiometric relationship in the stoichiometric balance sheet. The temperature response rate of each reaction zone is monitored; the heating power of the fast response zone is set to the standard value, and the heating power of the slow response zone is set to 1.5 times the standard value. The execution effect of each coordinated control is recorded, including the time to reach the temperature target, the concentration matching accuracy, and the total energy consumption. When the stoichiometric balance sheet shows a deficiency of a certain reactant, the heating power of that reaction is reduced by 30% to avoid energy waste. When the stoichiometric balance sheet shows an excess of reaction products, the heating power of the relevant reaction is appropriately increased by 20% to accelerate product conversion. By setting the linkage parameters between temperature and concentration, the execution order of the thermal management sequence and the addition sequence is determined, achieving sequential execution of temperature control and material addition.

[0063] Step S150: Decompose the cooperative control sequence into a fast response control loop and a slow response control loop. Configure PID parameters for the fast response control loop to generate a transient adjustment signal. Use the transient adjustment signal to drive the slow response control loop to form a cascade adjustment sequence. Use the cascade adjustment sequence to generate the final purification command.

[0064] Specifically, the coordinated control sequence is decomposed into fast-response control loops and slow-response control loops. The coordinated control sequence includes control information such as temperature commands, dosing commands, execution timing, and coordination conditions. These controlled objects have different response speeds and time constants. Based on the dynamic response characteristics of each component of the purification device, the controlled objects are divided into fast-response and slow-response categories. Fast-response control loops include electric heater power regulation, fan speed control, electric valve opening regulation, and spray pump frequency control, with response times in the second range. Slow-response control loops include washing liquid temperature control, catalyst bed temperature control, absorbent concentration regulation, and chemical dosing control, with response times ranging from minutes to hours. For example, when the SO2 concentration suddenly increases, the fast-response control loop can adjust the spray flow rate and fan speed within 10-30 seconds, while the slow-response control loop requires 5-15 minutes to adjust the liquid phase concentration by adding fresh absorbent. Control cycles are set for the fast-response control loop and the slow-response control loop respectively. The control cycle for the fast-response control loop is 1-5 seconds, and the control cycle for the slow-response control loop is 1-10 minutes. Control commands in the cooperative control sequence are assigned to the corresponding control loops according to their response characteristics to ensure that the commands of each control loop match its response capability.

[0065] PID parameters are configured for the fast-response control loop to generate transient control signals. The fast-response control loop requires rapid response and precise control; the PID control strategy can effectively handle transient disturbances and steady-state errors. The PID control output calculation formula is u(t) = Kp·e(t) + Ki·∫e(t)dt + Kd·de(t) / dt, where u(t) is the control output, Kp is the proportional gain, Ki is the integral gain, Kd is the derivative gain, and e(t) is the control error. PID parameters are determined for different types of fast-response control loops: a larger proportional gain and a moderate integral gain are used for the electric heater control; a medium proportional gain and a small derivative gain are used for the fan control; and a small proportional gain and a larger integral gain are used for the valve control. Step response tests are used to determine the dynamic characteristic parameters of each control loop, including time constant, overshoot, and settling time. Based on the dynamic characteristic parameters, parameter tuning methods are used to determine the optimal PID parameter combination, giving the control system good speed and stability. The output of the PID controller is limited within the safe operating range of the equipment to avoid overload or damage to the actuators. Transient control signals contain information such as control commands, execution time, output limits, and priorities, enabling timely response and precise adjustment to rapidly changing operating conditions.

[0066] In some embodiments, driving the slow response control loop to form a cascaded adjustment sequence via the transient adjustment signal includes: extracting an overshoot from the transient adjustment signal; generating a reverse compensation parameter based on the overshoot; modulating the slow response control loop using the reverse compensation parameter to generate a compensation output; and generating a cascaded adjustment sequence based on the alternating action of the transient adjustment signal and the compensation output.

[0067] Overshoot is extracted from transient control signals. Overshoot characteristics in the control process are extracted using signal analysis methods from the generated transient control signals. When responding to changes in the setpoint, transient control signals often exhibit overshoot exceeding the setpoint due to system inertia and control parameter settings. The overshoot is defined as σ = (ymax - yss) / yss × 100%, where σ is the overshoot, ymax is the maximum value during the response, and yss is the steady-state value. The time response curve of the transient control signal is analyzed to identify the peak point and steady-state value. The overshoot of each control channel is calculated, including temperature control overshoot, flow control overshoot, pressure control overshoot, and concentration control overshoot. The time characteristics of the overshoot are analyzed, including the time of overshoot occurrence, the duration of overshoot, and the rate of overshoot decay. The distribution characteristics of the overshoot are statistically analyzed, and the average overshoot, maximum overshoot, and standard deviation of the overshoot are calculated. The main causes of overshoot are identified, including excessive proportional gain, excessively short integral time, inappropriate derivative time, and large system inertia. Overshoot under different operating conditions is compared and analyzed to identify the correlation between overshoot and operating parameters. The extracted overshoot results include detailed information such as magnitude, occurrence time, duration, and degree of impact.

[0068] Reverse compensation parameters are generated based on overshoot. Overshoot reflects the dynamic performance deviation of the control system and requires reverse compensation to improve control quality. The reverse compensation parameter Cp is calculated based on the magnitude of the overshoot, and the compensation intensity is determined using a linear proportional relationship. Control loops with large overshoot correspond to larger compensation parameters, while control loops with small overshoot correspond to smaller compensation parameters. The average and maximum overshoot of each control loop are calculated using statistical analysis methods, serving as the basis for setting the compensation parameters. Corresponding compensation parameter values ​​are calculated for temperature control overshoot, flow control overshoot, and pressure control overshoot. The direction of the compensation parameter is opposite to the direction of the overshoot; a positive overshoot results in a negative effect, and a negative overshoot results in a positive effect. Considering the constraints of system stability, the range of compensation parameter values ​​is limited to a reasonable range to avoid control oscillations caused by overcompensation. The relationship between the compensation parameter and control performance is analyzed to determine the selection criteria for the optimal compensation intensity.

[0069] The slow-response control loop is modulated using a reverse compensation parameter to generate a compensated output. The calculated reverse compensation parameter Cp is used to numerically modulate the original output of the slow-response control loop. Mathematical operations are performed between the original output of the slow-response control loop and the reverse compensation parameter to generate a new output after compensation modulation. Reverse compensation modulation uses a subtraction operation, subtracting or adding the compensation parameter to the original output to suppress overshoot. For the washing liquid temperature control loop, the heating power output is modulated using the compensation parameter corresponding to temperature overshoot. For the catalyst bed temperature control loop, the temperature setpoint is modulated using the compensation parameter corresponding to heating overshoot. For the absorbent dosing control loop, the flow output command is modulated using the compensation parameter corresponding to dosing overshoot. The compensation modulation process considers the time delay characteristics of the slow-response control loop to ensure that the compensation effect matches the control response time. Upper and lower limits are set for the compensated output to ensure that the modulated output is within the safe operating range of the equipment. Compensation modulation uses a continuous calculation mode, updating the compensation parameter value based on real-time overshoot detection results. The modulated compensated output is directly transmitted to the actuator of the slow-response control loop to achieve precise control adjustment.

[0070] A cascaded control sequence is generated based on the alternating action of transient control signals and compensation outputs. The transient control signal is responsible for rapid response, while the compensation output is responsible for long-term stability. Their alternation achieves a balance between dynamic performance and steady-state accuracy. The timing of the alternation is determined based on the type and intensity of the disturbance; for small disturbances, the transient control signal dominates, while for large disturbances, the compensation output is enhanced. A switching condition is established: when the system deviation is less than a set threshold, fine-tuning is primarily performed using the transient control signal. When the system deviation exceeds the set threshold, the compensation output is activated for enhanced control. The alternation uses a weighted allocation method, with the total control output being utotal(t) = w1·ufast(t) + w2·uc(t), where w1 and w2 are weighting coefficients, ufast(t) is the transient control signal, and uc(t) is the compensation output. The weighting coefficients are dynamically adjusted according to the system state; w1 is larger under normal operating conditions, and w2 increases under abnormal operating conditions. A smooth transition mechanism for the alternation is established to avoid abrupt changes and oscillations in the control output. The cascade control sequence records the complete process of alternating action, including information such as signal type, action strength, switching time, and control effect. The cascade control sequence achieves an organic combination of rapid inner-loop regulation and stable outer-loop control, improving the dynamic performance and steady-state accuracy of the overall control system.

[0071] The final purification command is generated using a cascade control sequence. This sequence is then converted into specific equipment operating parameters, including heater power setpoints, fan speed setpoints, and dosing pump flow rate setpoints. Control commands are prioritized according to safety level: emergency shutdown commands are set to level 1 priority, equipment protection commands to level 2 priority, and routine adjustment commands to level 3 priority. When the scrubbing tower receives both heating and dosing commands simultaneously, the dosing command is executed first, followed by a 30-second wait before the heating command to avoid temperature shock. When the catalytic reactor receives both temperature and flow rate adjustment commands, the flow rate is adjusted to the set value before adjusting the temperature to ensure reaction stability. Execution intervals for each equipment command are set: 10 seconds for electric heater commands, 15 seconds for fan speed control commands, and 20 seconds for dosing pump adjustment commands. Control commands in the cascade control sequence are prioritized: safety protection commands have the highest priority, fast-response adjustment commands have a higher priority, and slow-response adjustment commands have a normal priority. The system monitors the equipment response time after command execution: 5-10 seconds for electric heaters, 10-20 seconds for fans, and 15-30 seconds for dosing pumps. It records the success rate and reasons for failure of command execution; successfully executed commands return an acknowledgment signal, while failed commands return an error code. Control commands are categorized and organized according to the executing equipment. Multiple commands for the same equipment require coordination and optimization to avoid command conflicts and mutual interference. Control commands are transmitted via digital signals, with each command containing basic information such as equipment number, operation type, and parameter values. Ultimately, the purification commands cover all purification equipment, including scrubbing towers, catalytic reactors, and bag filters, forming a complete system-level control scheme.

[0072] To implement the ship exhaust gas dynamic purification method corresponding to the above method embodiments, and to achieve the corresponding functions and technical effects. See also Figure 2 , Figure 2 This diagram illustrates a structural block diagram of a dynamic purification device 200 for ship exhaust gas according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The dynamic purification device 200 for ship exhaust gas according to this embodiment includes: Data acquisition module 201 is used to collect multi-point sensor data of ship exhaust gas and operating parameters of purification device. The multi-point sensor data includes pollutant concentration distribution and temperature field distribution. The multi-point sensor data is used to extract time-domain features to construct a real-time monitoring network. A closed-loop control loop is established based on the real-time monitoring network and the operating parameters of the purification device. Bottleneck identification module 202 is used to identify the purification bottleneck area by mass transfer resistance analysis based on the closed-loop control loop, implement a dense sensor arrangement in the purification bottleneck area to generate a dense monitoring field, collect monitoring data from the dense monitoring field, perform disturbance compensation to form a steady-state measurement value, and construct an adaptive control domain through the steady-state measurement value; The feedforward generation module 203 is used to evaluate the absorbent activity of the adaptive control domain, extract the attenuation characteristics, back-map the attenuation characteristics to generate a feedforward control signal for supplementary dosage, use the feedforward control signal to trigger precise dosing to form a concentration matching curve, and establish a stoichiometric balance table based on the concentration matching curve. Temperature control compensation module 204 is used to perform micro-area reaction kinetic analysis and identify rapid control steps through the encrypted monitoring field, apply temperature regulation to the rapid control steps to generate activation energy compensation, convert the activation energy compensation into heating power commands to form a thermal management sequence, and couple the thermal management sequence with the stoichiometric balance table to generate a collaborative control sequence. The cascade control module 205 is used to decompose the cooperative control sequence into a fast-response control loop and a slow-response control loop, configure PID parameters for the fast-response control loop to generate a transient control signal, drive the slow-response control loop to form a cascade control sequence through the transient control signal, and use the cascade control sequence to generate the final purification control command.

[0073] The aforementioned ship exhaust gas dynamic purification device 200 can implement the ship exhaust gas dynamic purification method of the above-described method embodiments. The options in the above method embodiments are also applicable to this embodiment, and will not be detailed here. The remaining contents of this application embodiment can be referred to the contents of the above method embodiments, and will not be repeated in this embodiment.

[0074] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A method for dynamic purification of ship exhaust gas, characterized in that, include: Collect multi-point sensor data of ship exhaust gas and operating parameters of purification device. The multi-point sensor data includes pollutant concentration distribution and temperature field distribution. Extract time-domain features from the multi-point sensor data to construct a real-time monitoring network. Establish a closed-loop control loop based on the real-time monitoring network and the operating parameters of the purification device. Mass transfer resistance analysis is performed based on the closed-loop control loop to identify the purification bottleneck area. Sensors are densified in the purification bottleneck area to generate a densified monitoring field. Monitoring data is collected from the densified monitoring field and disturbance compensation is performed to form steady-state measurement values. An adaptive control domain is constructed using the steady-state measurement values. The absorbent activity of the adaptive control domain is evaluated to extract attenuation characteristics. The attenuation characteristics are then mapped back to the supplementary dosage to generate a feedforward control signal. The feedforward control signal is used to trigger precise dosing to form a concentration matching curve. A stoichiometric balance table is then established based on the concentration matching curve. Micro-region reaction kinetics analysis is performed through the encrypted monitoring field to identify rate control steps. Temperature regulation is applied to the rate control steps to generate activation energy compensation. The activation energy compensation is converted into heating power commands to form a thermal management sequence. Based on the thermal management sequence and the stoichiometric balance table, a collaborative control sequence is generated. The coordinated control sequence is decomposed into a fast-response control loop and a slow-response control loop. PID parameters are configured for the fast-response control loop to generate a transient adjustment signal. The transient adjustment signal drives the slow-response control loop to form a cascade adjustment sequence. The cascade adjustment sequence is then used to generate the final purification control command.

2. The method according to claim 1, characterized in that, The step of implementing a encrypted sensor deployment to generate an encrypted monitoring field in the purification bottleneck area includes: A swing correction distribution map is generated using the aforementioned purification bottleneck area; A monitoring point-correction correspondence is formed based on the swing correction distribution map; High-gradient monitoring areas are selected from the monitoring point-correction correspondence; An encrypted monitoring field is formed by the encrypted deployment of the high-gradient monitoring area.

3. The method according to claim 1, characterized in that, The step of collecting monitoring data from the encrypted monitoring field, performing disturbance compensation, and forming steady-state measurement values ​​includes: The encrypted monitoring field data is collected and segmented into mainstream signals and disturbance signals; Extract phase difference features between the mainstream signal and the disturbance signal; A time delay compensation matrix is ​​constructed using the phase difference characteristics; The disturbance signal is phase-corrected based on the time delay compensation matrix to generate steady-state measurement values.

4. The method according to claim 1, characterized in that, The step of inversely mapping the attenuation characteristics to generate a feedforward control signal based on the supplementary dosage includes: The attenuation characteristics are used to identify the activity loss rate and generate a loss gradient; Based on the loss gradient, multiple supplementary intervals are divided to form a tiered demand spectrum; The segmented supplementary sequence is generated by matching the differentiated investment acceleration rate with the tiered demand spectrum. The segmented supplementary sequence is configured with a trigger threshold to generate a feedforward control signal.

5. The method according to claim 1, characterized in that, The step of identifying rate control through micro-region reaction kinetic analysis using the encrypted monitoring field includes: The encrypted monitoring field data is converted into a micro-region reaction rate field; Identify reaction stagnation regions from the micro-region reaction rate field; The reaction stagnation region is subjected to boundary perturbation injection to form an activation path; The speed control step is formed by locating the speed limiting link along the activation path.

6. The method according to claim 1, characterized in that, The step of applying temperature control to generate activation energy compensation in the rate control step includes: The reaction barrier distribution is constructed using the rate-controlling step described above; Energy analysis is performed on the reaction barrier distribution to generate barrier height markers; The heating zone is determined using the aforementioned barrier height marker; Activation energy compensation is generated by increasing the temperature in the heating zone.

7. The method according to claim 1, characterized in that, The step of driving the slow-response control loop to form a cascaded regulation sequence through the transient adjustment signal includes: Extract the overshoot from the transient adjustment signal; Reverse compensation parameters are generated based on the overshoot. The slow response control loop is modulated using the inverse compensation parameters to generate a compensation output. A cascaded adjustment sequence is generated based on the alternating action of the transient adjustment signal and the compensation output.

8. The method according to claim 2, characterized in that, The formation of an encrypted monitoring field through the encrypted deployment of the high-gradient monitoring area includes: Identify flow field dead zones based on the high gradient monitoring region; An activation node is set at the boundary of the dead zone in the flow field; The effective monitoring range is formed by collecting dead zone revival signals from the activated nodes. An encrypted monitoring field is generated by adjusting the complementary relationship of the effective monitoring range.

9. The method according to claim 6, characterized in that, The activation energy compensation generated by the temperature increase in the heating zone includes: The temperature sequence of the heating zone is converted into a temperature gradient distribution; The temperature gradient distribution drives the gas-liquid interface to update, generating an increment in the mass transfer coefficient. Extract the reaction acceleration factor from the mass transfer coefficient increment; Activation energy compensation is generated according to the cumulative effect of the reaction acceleration factor.

10. A dynamic purification device for ship exhaust gas, characterized in that, include: The data acquisition module is used to collect multi-point sensor data of ship exhaust gas and operating parameters of the purification device. The multi-point sensor data includes pollutant concentration distribution and temperature field distribution. The module extracts time-domain features from the multi-point sensor data to construct a real-time monitoring network. Based on the real-time monitoring network and the operating parameters of the purification device, a closed-loop control loop is established. The bottleneck identification module is used to identify the purification bottleneck area by mass transfer resistance analysis based on the closed-loop control loop, implement a dense sensor arrangement in the purification bottleneck area to generate a dense monitoring field, collect monitoring data from the dense monitoring field, perform disturbance compensation to form a steady-state measurement value, and construct an adaptive control domain through the steady-state measurement value. The feedforward generation module is used to evaluate the absorbent activity of the adaptive control domain, extract the attenuation characteristics, back-map the attenuation characteristics to generate a feedforward control signal for supplementary dosage, use the feedforward control signal to trigger precise dosing to form a concentration matching curve, and establish a stoichiometric balance table based on the concentration matching curve. The temperature control compensation module is used to identify the rapid control step through micro-region reaction kinetic analysis in the encrypted monitoring field, apply temperature regulation to the rapid control step to generate activation energy compensation, convert the activation energy compensation into heating power command to form a thermal management sequence, and couple the thermal management sequence with the stoichiometric balance table to generate a collaborative control sequence. The cascade control module is used to decompose the cooperative control sequence into a fast-response control loop and a slow-response control loop, configure PID parameters for the fast-response control loop to generate a transient control signal, drive the slow-response control loop to form a cascade control sequence through the transient control signal, and use the cascade control sequence to generate the final purification control command.

Images