Intelligent control method and device for carbon capture of a ship

Abstract

The application discloses a ship carbon capture intelligent regulation and control method and device, which detects abnormal fluctuation of temperature, pressure and flow parameters of a carbon capture device, identifies deviated data segments, and performs fault feature analysis to generate a diagnosis mark sequence; based on the diagnosis mark sequence, a health state flow is generated by reconstructing the running state, a fuzzy control signal is generated by extracting a control membership degree through fuzzy processing; a multivariate regulation and control channel is established to analyze the absorption efficiency, implement the lean-liquid circulation optimization to form a strengthened absorption domain; a load distribution is executed to generate a parallel capture flow of a tower, solvent degradation monitoring and pollution source positioning analysis are performed to generate a solvent update timing; the capture energy consumption is step-adjusted by using the solvent update timing, peak-valley scheduling optimization is implemented based on a capture capacity curve to determine a switching control time, and working condition adaptive adjustment is performed at the switching control time, so that intelligent regulation and control and optimized operation of the ship carbon capture device are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental pollution control, in particular to a ship carbon capture intelligent regulation method and device. BACKGROUND

[0002] With the increasingly stringent standards of the International Maritime Organization on ship carbon emissions, ship carbon capture technology is facing unprecedented technical challenges. The traditional ship carbon capture system mainly relies on fixed parameter operation and manual experience adjustment, and has problems such as large fluctuation of processing efficiency, high equipment failure rate and high operation cost in complex marine environments. The existing ship carbon capture device generally adopts simple on-off control and timing maintenance strategy, lacks real-time grasp and prediction ability of the internal working state of the system, and has obvious shortcomings in equipment health management, process parameter coordination and load change adaptation, which is difficult to meet the requirements of stable and efficient operation of the carbon capture system under different navigation conditions of the ship.

[0003] Therefore, it is urgent to develop an intelligent regulation technology to solve at least one of the above problems. SUMMARY

[0004] The present application discloses a ship carbon capture intelligent regulation method and device, which aims to realize system health management through real-time monitoring and fault warning of equipment state, improve regulation accuracy by using fuzzy reasoning and multi-parameter coordination technology, realize efficient and stable operation by combining process optimization and load scheduling, and ensure the optimal performance of the system under various working conditions through prediction analysis and adaptive adjustment.

[0005] The first aspect of the present application proposes a ship carbon capture intelligent regulation method, comprising the following steps:

[0006] Collecting temperature, pressure and flow parameters of the carbon capture device, identifying deviated data segments by detecting abnormal fluctuations of the temperature, pressure and flow parameters, and generating a diagnosis marker sequence by analyzing fault features using the deviated data segments;

[0007] Reconstructing the running state based on the diagnosis marker sequence to generate a health state stream, performing fuzzy processing on the health state stream to extract control membership, generating a fuzzy control signal by using the control membership to stimulate a pre-set reasoning rule library, and establishing a multivariable regulation channel through the fuzzy control signal;

[0008] Performing absorption efficiency analysis on the multivariable regulation channel to generate solvent performance parameters, implementing lean-rich liquid circulation optimization based on the solvent performance parameters to form a strengthened absorption domain, and generating a parallel capture stream by performing load distribution on the strengthened absorption domain;

[0009] The solvent degradation monitoring of the parallel capture flow of the sub-tower identifies the inactivation distribution characteristics, the pollution source positioning analysis is carried out based on the inactivation distribution characteristics, the solvent updating time sequence is generated, the energy consumption of the capture is step-adjusted by using the solvent updating time sequence, and the energy-saving operation configuration is constructed;

[0010] The carbon load analysis is carried out based on the energy-saving operation configuration, the capture capacity curve is generated, the peak-valley scheduling optimization is implemented on the capture capacity curve, the switching control time is determined, the working condition adaptive adjustment is executed at the switching control time, and the intelligent control of carbon capture is completed.

[0011] The second aspect of the present application provides a ship carbon capture intelligent regulation and control device, comprising:

[0012] The fault diagnosis module is used for collecting the temperature, pressure and flow parameters of the carbon capture device, detecting abnormal fluctuations of the temperature, pressure and flow parameters to identify the deviated data segment, and generating a diagnosis mark sequence by using the deviated data segment for fault feature analysis;

[0013] The fuzzy control module is used for generating a health state flow by reconstructing the running state based on the diagnosis mark sequence, implementing fuzzy processing on the health state flow to extract a control membership degree, generating a fuzzy control signal by using the control membership degree to stimulate a preset reasoning rule library, and establishing a multivariable regulation and control channel through the fuzzy control signal;

[0014] The absorption optimization module is used for generating solvent performance parameters by analyzing the absorption efficiency of the multivariable regulation and control channel, implementing lean-rich liquid circulation optimization based on the solvent performance parameters to form a strengthened absorption domain, and generating a parallel capture flow of sub-tower by performing load distribution on the strengthened absorption domain;

[0015] The regeneration management module is used for monitoring the solvent degradation of the parallel capture flow of the sub-tower to identify the inactivation distribution characteristics, performing pollution source positioning analysis based on the inactivation distribution characteristics to generate a solvent updating time sequence, and step-adjusting the energy consumption of the capture by using the solvent updating time sequence to construct an energy-saving operation configuration;

[0016] The intelligent scheduling module is used for generating a capture capacity curve by analyzing the carbon load based on the energy-saving operation configuration, implementing peak-valley scheduling optimization on the capture capacity curve, determining a switching control time, and executing working condition adaptive adjustment at the switching control time to complete the intelligent control of carbon capture.

[0017] The beneficial effects of the present application are embodied in the following points: first, by establishing a perfect device health monitoring and fault warning mechanism, the abnormal state and potential risk in system operation can be found in time, the traditional periodic maintenance is changed into state-based preventive maintenance, the device failure rate and maintenance cost are effectively reduced, and the reliability of system operation is improved. Secondly, through the application of fuzzy control and multivariable coordination technology, intelligent coordination of key parameters such as solvent circulation, temperature control and load distribution is realized, the problems of improper parameter matching and response lag in traditional manual adjustment are solved, the system can quickly adapt to the working condition change and maintain stable capture efficiency. Finally, through solvent performance analysis and update timing optimization, a solvent management strategy based on actual demand is established, the waste of solvent and performance loss caused by traditional regular replacement are avoided, combined with peak-valley scheduling of capture capacity, effective control of system energy consumption is realized, and the operation cost of ship carbon capture is reduced.

[0018] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0019] The drawings herein show specific examples of the technical solutions described in the present application, and constitute part of the specification together with the specific embodiments, for explaining the technical solutions, principles and effects of the present application.

[0020] Unless specifically stated, the same reference signs in different drawings represent the same or similar technical features, and different reference signs may also be used to represent the same or similar technical features.

[0021] Figure 1 is a flowchart of a ship carbon capture intelligent control method of the present application.

[0022] Figure 2 is a structural block diagram of a ship carbon capture intelligent control device of the present application. DETAILED DESCRIPTION

[0023] In the following description, specific details are set forth in order to provide a thorough understanding of the present application. However, persons skilled in the art will understand that the present application can be practiced without these specific details. In other instances, well-known structures, devices, circuits and methods have not been described in detail in order to avoid obscuring the present application.

[0024] 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.

[0025] The technical solutions of the embodiments of this application are described below.

[0026] like Figure 1 As shown, this embodiment of the invention provides a smart control method for ship carbon capture, including the following steps S110-S150:

[0027] Step S110: Collect the temperature, pressure and flow parameters of the carbon capture device, detect and identify abnormal fluctuations in the temperature, pressure and flow parameters to identify deviation data segments, and use the deviation data segments to perform fault feature analysis and generate a diagnostic marker sequence.

[0028] Specifically, temperature, pressure, and flow parameters of the carbon capture unit are collected. High-precision sensor arrays are deployed at key nodes of the carbon capture unit to monitor changes in temperature, pressure, and flow parameters in real time during operation. Temperature parameter monitoring covers the internal temperature distribution of the absorption tower, the heating temperature of the regeneration tower, the outlet temperature of the cooler, and the temperatures between each stage of the compressor. Pt100 resistance temperature sensors and thermocouple sensors are used, achieving a measurement accuracy of ±0.1℃ and a response time of less than 5 seconds. Pressure parameter monitoring includes the operating pressure of the absorption tower, the steam pressure of the regeneration tower, the pressures at each stage of the CO2 compressor, and the pipeline delivery pressure. Diffused silicon pressure sensors and strain gauge pressure transmitters are used, covering a measurement range of 0-10 MPa with an accuracy class of 0.25%. Flow parameter monitoring involves the amine liquid circulation flow rate, steam supply flow rate, cooling water flow rate, and CO2 product flow rate. Electromagnetic flow meters, vortex flow meters, and mass flow meters are used, achieving a measurement accuracy better than ±1%. The sensors are systematically arranged according to the series and parallel relationships of the process flow, forming monitoring sub-networks in the absorption, regeneration, compression, and purification sections. The data acquisition system adopts a distributed architecture. The field acquisition units communicate with the central control system via Modbus protocol and Ethernet. The acquisition frequency is set to 1Hz for continuous sampling to ensure the capture of dynamic changes in temperature, pressure and flow parameters.

[0029] The abnormal fluctuation of temperature, pressure and flow parameters is detected to identify the deviation data segment. Based on the collected time series data of temperature, pressure and flow parameters, the abnormal fluctuation mode of the parameters is detected through statistical analysis and signal processing method. The statistical characteristic benchmark of normal operation is established for the temperature, pressure and flow parameters, the mean value μ, the standard deviation σ and the variation coefficient CV = σ / μ of each parameter are calculated, and the abnormal detection threshold of 3σ criterion is established. The sliding window technology is used for real-time abnormal detection of the temperature, pressure and flow parameters, the window length is set to 300 sampling points, the sliding step is 30 sampling points, and the statistical characteristics of the parameters are calculated in each window. When the parameter value exceeds the range of μ±3σ, it is marked as an abnormal point, and when more than 5 consecutive sampling points are abnormal, it is determined as an abnormal fluctuation segment. The wavelet transform is used for multi-scale decomposition of the temperature, pressure and flow parameters, and the abnormal characteristics of different frequency components are identified. High-frequency abnormality usually corresponds to equipment failure, medium-frequency abnormality corresponds to process disturbance, and low-frequency abnormality corresponds to load change. The correlation analysis between parameters is established, when temperature abnormality is accompanied by pressure abnormality, equipment performance abnormality is identified first; when flow abnormality appears independently, control system abnormality is identified. The detected abnormal fluctuations are classified and marked, including sudden abnormality, gradual abnormality, periodic abnormality and random abnormality. The starting time, ending time, peak value, duration and influence range of each abnormal fluctuation segment are recorded, and the structured description of the deviation data segment is formed. The deviation data segment contains key attribute information such as abnormal type, severity, time characteristics and spatial distribution.

[0030] In some embodiments, the fault feature analysis using the deviation data segment generates a diagnostic marker sequence, including: extracting an abnormal amplitude distribution according to the deviation data segment; applying the abnormal amplitude distribution to time correlation to identify a fault evolution path; constructing a failure risk level based on the fault evolution path; and preparing a diagnostic marker sequence according to the failure risk level.

[0031] Extract abnormal amplitude distribution from deviated data segments. Based on the identified deviated data segments, the probability distribution characteristics of abnormal amplitude are extracted by statistical analysis method. The deviated data segments contain the amplitude information of parameter deviation from normal value, and the abnormal amplitude distribution reflects the intensity characteristics and randomness characteristics of fault. For each deviated data segment, the abnormal amplitude A = |X-X_normal| is calculated, where X is the parameter value at the abnormal time, and X_normal is the corresponding normal value. The abnormal amplitudes of all deviated data segments are counted to construct the frequency histogram of abnormal amplitude, and the concentration trend and dispersion degree of amplitude distribution are analyzed. The statistical parameters of abnormal amplitude distribution are calculated, including mean, variance, skewness and kurtosis, which reflect the typical intensity and distribution form of abnormality. The optimal distribution type of abnormal amplitude is determined using probability distribution fitting method, and the common distributions include normal distribution, exponential distribution, Weibull distribution and gamma distribution. The accuracy of distribution fitting is verified by K-S test and chi-square test, and the distribution with the best fitting effect is selected as the probability distribution of abnormal amplitude. The amplitude distribution of different types of abnormality is extracted respectively, and temperature abnormality, pressure abnormality and flow abnormality have different amplitude distribution characteristics. The parameter table of abnormal amplitude distribution is established to record the distribution type, distribution parameter, goodness of fit and confidence interval, etc. The abnormal amplitude distribution provides quantitative abnormal intensity information for subsequent fault evolution analysis.

[0032] Apply time correlation to identify fault evolution path based on abnormal amplitude distribution. Abnormal amplitude distribution describes the intensity characteristics of abnormality, and time correlation analysis reveals the dynamic process of abnormal development. The abnormal amplitudes are arranged in chronological order to construct the time series A(t) of abnormal amplitude. The change rate dA / dt of adjacent time abnormal amplitude is calculated, where dA is the abnormal amplitude change, dt is the time interval, positive value indicates that the abnormality is aggravated, and negative value indicates that the abnormality is alleviated. The moving average method is used to smooth the abnormal amplitude time series to eliminate the influence of random fluctuations and highlight the long-term trend. The development mode of abnormal amplitude is identified through trend analysis, including rising trend, falling trend, stable trend and fluctuating trend. The correlation relationship between abnormal events is established, and when the interval between the end time of the current abnormal event and the start time of the next abnormal event is less than the set threshold, it is considered that the two events have correlation relationship. The sequence pattern of abnormal events is identified through association rule mining to find typical fault evolution sequences. The fault evolution path graph is constructed, with abnormal events as nodes and time correlation as edges, forming a directed graph structure. The transition probability of fault evolution path is calculated to quantify the possibility of turning from one abnormal state to another. The fault evolution path records the complete process of fault occurrence, development, deterioration and recovery, including path nodes, transition probability, duration and evolution direction, etc.

[0033] Illustratively, the failure risk level is constructed based on the failure evolution path, including: tracking parameter mutation points through the failure evolution path; establishing correlation strength between the mutation points and quantifying it as a risk transmission coefficient; starting a warning mechanism when the risk transmission coefficient continues to rise, and maintaining a monitoring state when the risk transmission coefficient tends to be stable; and forming a failure risk level by integrating the trigger frequency of the warning mechanism and the duration of the monitoring state.

[0034] The parameter mutation points are tracked through the failure evolution path. The failure evolution path records the time sequence development process of abnormal events, and the parameter mutation point is a key node of fault state conversion. First-order difference calculation is performed on the parameter time sequence in the failure evolution path to obtain a parameter change rate sequence. The sliding variance method is used to detect the sharp fluctuations of the parameter change rate, and when the local variance exceeds 3 times the global variance, it is marked as a potential mutation point. The CUSUM control chart method is used for precise positioning of the mutation point, and the cumulative sum sequence C(t)=∑[X(i)-μ] is calculated, where X(i) is the i-th parameter value and μ is the parameter mean. When the cumulative sum sequence jumps obviously, the corresponding time point is the parameter mutation point. The detected mutation points are classified, and according to the mutation direction, they are divided into positive mutation and negative mutation, and according to the mutation amplitude, they are divided into strong mutation and weak mutation. The occurrence frequency and distribution law of the mutation points are analyzed, high-frequency mutation represents system instability, and low-frequency mutation represents occasional abnormalities. The time position, mutation amplitude, mutation direction and influence parameter of each mutation point are recorded to form the attribute description of the mutation point. The corresponding relationship between the mutation point and the failure type is established, and different failures have characteristic mutation modes. Through the mutation point tracking, the key turning points in the failure evolution process are accurately identified and positioned.

[0035] The correlation strength between the mutation points is established and quantified as a risk transfer coefficient. By tracking the parameter mutation points, the correlation strength between the mutation points is established by correlation analysis, and the correlation strength is quantified as a risk transfer coefficient. The correlation strength between adjacent mutation points reflects the propagation characteristics and influence degree of the fault in the system. The time interval Δt and the amplitude difference ΔA of adjacent mutation points are calculated, and the shorter the time interval and the greater the amplitude difference, the stronger the correlation strength. The cross-correlation function is used to analyze the correlation between different parameter mutation points, and the mutation point pair with a cross-correlation coefficient greater than 0.7 is considered to have strong correlation. The calculation formula of the risk transfer coefficient K is established, K=(ΔA2 / ΔA1)×exp(-Δt / τ), where K is the risk transfer coefficient, ΔA1 and ΔA2 are the amplitudes of the previous and subsequent mutation points, and τ is the time constant of the system. The risk transfer coefficient K>1 represents risk amplification, K=1 represents risk translation, and K<1 represents risk attenuation. The chain correlation analysis is performed on multiple mutation points, and the total transfer coefficient K_total of the risk transfer chain is calculated. The trend of the risk transfer coefficient is analyzed, and an upward trend indicates fault aggravation, and a downward trend indicates fault mitigation. A threshold system of the risk transfer coefficient is established, and a key control point and a warning point are set. The time series of the risk transfer coefficient is recorded to form dynamic monitoring data of the risk transfer.

[0036] When the risk transfer coefficient continues to rise, the warning mechanism is started, and when the risk transfer coefficient tends to be stable, the monitoring state is maintained. Based on the calculated risk transfer coefficient trend, dynamic risk management is achieved through state judgment and mechanism switching. The trend of the risk transfer coefficient reflects the development direction and severity of the system risk. The change slope of the risk transfer coefficient is calculated using linear regression method, a positive slope indicates risk increase, a negative slope indicates risk decrease, and a zero slope indicates risk stability. The rising threshold α and the duration threshold T_warn of the risk transfer coefficient are set, and when the slope is greater than α and the duration exceeds T_warn, the warning mechanism is started. The warning mechanism includes multiple levels of warning methods such as sound and light alarm, short message notification, email reminder and emergency response. When the change slope of the risk transfer coefficient is less than the stability threshold β and the change amplitude is less than the tolerance range ε, the system enters the monitoring state. In the monitoring state, the system maintains normal data acquisition and analysis frequency, and does not trigger additional warning and response measures. The judgment logic of state switching is established, and the risk increase condition needs to be met from the monitoring state to the warning state, and the risk stability condition needs to be met from the warning state to the monitoring state. The trigger time, duration and trigger reason of the warning mechanism are recorded to form the historical record of the warning event. The entry time, duration and state characteristics of the monitoring state are recorded to form the normal record of the system operation.

[0037] The trigger frequency of the comprehensive early warning mechanism and the duration of the monitoring state form the failure risk level. The number of triggers N_alarm and the total duration T_alarm of the statistical early warning mechanism in the evaluation period are counted. The total duration T_monitor of the monitoring state and the longest continuous monitoring time T_max in the evaluation period are counted. The warning frequency F_alarm=N_alarm / T_total and the monitoring ratio R_monitor=T_monitor / T_total are calculated, where T_total is the total duration of the evaluation period. A risk level evaluation matrix is established, with the horizontal axis as the warning frequency and the vertical axis as the monitoring ratio. Different regions in the matrix correspond to different risk levels. When F_alarm>0.3 and R_monitor<0.5, the failure risk level is extremely high risk; when F_alarm>0.2 and R_monitor<0.7, the failure risk level is high risk; when F_alarm>0.1 and R_monitor<0.8, the failure risk level is medium risk; when F_alarm<0.1 and R_monitor>0.8, the failure risk level is low risk. Considering the weighted influence of warning severity, the weight coefficient of high severity warning is 2, the weight coefficient of medium severity warning is 1, and the weight coefficient of low severity warning is 0.5. The failure risk level is an important indicator of equipment health management, guiding maintenance decision and resource allocation, and realizing risk-based preventive maintenance strategy.

[0038] According to the failure risk level, a diagnosis mark sequence is prepared. A unique mark code is assigned to each risk level, with low risk marked as L, medium risk marked as M, high risk marked as H, and extremely high risk marked as C. Device number, fault type and time information are added to the mark code to form a complete diagnosis mark format, such as "H-T01-P05-20240315" indicating that device T01 had a high-risk P05 fault on March 15, 2024. A priority order of diagnosis marks is established, with extremely high risk marks having the highest priority, high risk marks having high priority, medium risk marks having medium priority, and low risk marks having low priority. According to the risk level, the corresponding processing suggestions are matched, with low risk performing normal monitoring, medium risk increasing inspection frequency, high risk arranging planned maintenance, and extremely high risk immediately shutting down for maintenance. The diagnosis marks are arranged in time order and priority order to form a diagnosis mark sequence. The diagnosis mark sequence is stored in a queue structure, supporting dynamic addition and deletion operations. Each diagnosis mark contains complete fault information, risk level, processing suggestion and execution state, etc. The diagnosis mark sequence provides systematic decision support for equipment maintenance management, realizing preventive maintenance and risk control of faults.

[0039] In step S120, a health state flow is generated based on the running state reconstruction of the diagnostic marker sequence, a fuzzy processing is performed on the health state flow to extract a control membership degree, a fuzzy control signal is generated by using the control membership degree to stimulate a preset reasoning rule base, and a multivariable regulation and control channel is established by using the fuzzy control signal.

[0040] Specifically, a health state flow is generated based on the running state reconstruction of the diagnostic marker sequence. The health state flow refers to converting the discrete diagnostic marker sequence into a continuous device health degree time sequence. The risk level in the diagnostic marker sequence is converted into a numerical health state value, with low risk corresponding to a health state value of 0.9-1.0, medium risk corresponding to 0.7-0.9, high risk corresponding to 0.4-0.7, and extremely high risk corresponding to 0.0-0.4. A health state time sequence is constructed for each device, and the discrete diagnostic markers are interpolated in time sequence to form a continuous health state curve. A spline interpolation method is used to process the state transition between diagnostic markers to ensure the smoothness and continuity of the health state flow. The influence range and propagation effect of the fault are considered, and when a fault occurs in a device, the health state of the associated devices will also be affected. For example, when the absorption tower has a solvent degradation fault, its health state value decreases from 0.95 to 0.6, and the health states of the regeneration tower and the compressor decrease to 0.8 and 0.85, respectively. The comprehensive health state H_sys of the system level is calculated by a weighted average method, where w_i is the device weight coefficient and H_i is the single device health state. The health state flow records the complete health evolution process of the carbon capture device from normal operation to fault occurrence and then to recovery, including state values, change trends, fluctuation characteristics, and correlation information.

[0041] The health state flow is fuzzified to extract control membership. The deterministic health state value is converted into fuzzy control membership by fuzzy set theory. The health state flow provides a numerical description of the health degree of the equipment, and the fuzzification process can handle the uncertainty and fuzziness of the health state boundary. The fuzzy set of health state is defined, including five fuzzy subsets of "excellent", "good", "general", "poor" and "bad". The trapezoidal membership function is used to describe each fuzzy subset. The membership function of the excellent state is 1 in the interval of 0.9-1.0, and linearly decreases to 0 in the interval of 0.8-0.9. The membership function of the good state is 1 in the interval of 0.7-0.9, and linearly changes in the intervals of 0.6-0.7 and 0.9-1.0. The membership of each state value in the health state flow to each fuzzy subset is calculated, μ_i(x)=f_i(H), where f_i is the membership function and H is the health state value. The influence of the trend of health state on the membership is considered. The upward trend increases the membership of "excellent" and "good", and the downward trend increases the membership of "poor" and "bad". The control membership vector μ=[μ_1,μ_2,μ_3,μ_4,μ_5] is calculated, and each element in the vector corresponds to the membership of a fuzzy subset. The control membership is calculated for different equipment and different parameters to form a multi-dimensional membership matrix. The control membership reflects the fuzzy characteristics and uncertainty of the health state of the equipment, and is the input variable for subsequent fuzzy reasoning.

[0042] In some embodiments, the generating the fuzzy control signal by triggering the preset inference rule base according to the control membership includes: obtaining an overlapping interval from the control membership; deriving a control weight distribution according to the overlapping interval; triggering a preset inference rule base to generate an execution priority based on the control weight distribution; and determining the fuzzy control signal by the execution priority.

[0043] The overlapping interval is obtained from the control membership. The control membership describes the degree of membership of the health state to each fuzzy subset, and the overlapping interval between adjacent fuzzy subsets reflects the fuzziness of state transition. The control membership vector μ = [μ_1, μ_2, μ_3, μ_4, μ_5] is analyzed to identify the combination of fuzzy subsets with non-zero membership at the same time. When the membership of two adjacent fuzzy subsets is greater than 0, there is an overlapping interval, and the overlapping strength is defined as the minimum of the two memberships O_ij = min(μ_i, μ_j). For example, when the health state belongs to "good" (membership 0.6) and "general" (membership 0.4) at the same time, the overlapping interval strength is 0.4. The distribution pattern of the overlapping interval is analyzed, single overlap indicates that two adjacent fuzzy subsets overlap, and multiple overlap indicates that three or more fuzzy subsets overlap at the same time. The coverage and impact of the overlapping interval are calculated, and the larger the coverage, the stronger the state fuzziness, which requires more cautious control strategy. The frequency of different overlapping patterns is counted, and the common overlapping patterns include excellent-good overlap, good-general overlap, general-poor overlap, and poor-evil overlap.

[0044] The control weight distribution is derived based on the overlapping interval. The overlapping interval reflects the fuzziness and uncertainty of the health state, and needs to be balanced by weight distribution to balance the strength of different control strategies. The normalized weight distribution method is used to convert the membership of the overlapping interval into weight coefficients w_i = μ_i / Σμ_j, where w_i is the weight of the i-th control strategy, and μ_i is the corresponding membership. The influence of the fuzzy degree of the overlapping interval on the weight distribution is considered, and the interval with high fuzzy degree adopts uniform weight distribution, and the interval with low fuzzy degree adopts concentrated weight distribution. The risk preference coefficient is introduced to adjust the weight distribution, and the conservative strategy increases the weight of safety control, and the aggressive strategy increases the weight of performance optimization. The comprehensive weight calculation is performed on multiple overlapping intervals, and the weighted average method W_total = Σ(O_i × w_i) is used, where W_total is the comprehensive weight, and O_i is the overlapping strength. The constraint condition of weight distribution is established to ensure that the sum of all weight coefficients is 1, and the single weight coefficient does not exceed the preset upper limit value. The stability and sensitivity of the weight distribution are analyzed, and the too frequent change of the weight will lead to control oscillation, which needs to use the smoothing filter method. The result of the control weight distribution determines the relative importance and impact of each control strategy in fuzzy reasoning.

[0045] The execution priority is generated based on the control weight distribution triggering a preset inference rule base. The preset inference rule base is pre-established based on the operation mechanism of the carbon capture device and expert experience, and adopts an IF-THEN form to store the mapping relationship between the health state and the control action. The rules in the preset inference rule base are sorted according to the corresponding weight coefficients, the rules with high weight coefficients have high priority, and the rules with low weight coefficients have low priority. A priority triggering threshold is set, only the rules with weight coefficients exceeding the threshold are triggered for execution, avoiding the control interference caused by low-weight rules. A hierarchical priority system is adopted, and the rule priority is divided into four levels: emergency level, high level, medium level and low level. The emergency level priority corresponds to the rules with weight coefficients greater than 0.8, which usually involves safety protection and emergency shutdown; the high level priority corresponds to the rules with weight coefficients of 0.6-0.8, which involves important parameter adjustment; the medium level priority corresponds to the rules with weight coefficients of 0.4-0.6, which involves performance optimization; and the low level priority corresponds to the rules with weight coefficients of 0.2-0.4, which involves auxiliary adjustment. Considering the conflict and competition relationship between the rules, when multiple rules have the same priority, rule coverage and rule credibility are used for secondary sorting. The activation history and execution effect of the rules are recorded, and the rules successfully executed obtain an additional reward coefficient in the weight distribution. The execution priority sequence is arranged in time sequence and priority order to form the execution queue of the inference rules.

[0046] The fuzzy control signal is determined by the execution priority. The inference rules are executed in turn according to the priority order, the high-priority rules are executed first, and the low-priority rules are executed after the high-priority rules are executed. The output contribution u_k of each executed rule is calculated, where u_k is the output contribution of the kth rule (control quantity unit), α_k is the activation strength, μ_k is the conclusion membership degree, and w_k is the weight coefficient. The weighted aggregation method is used to synthesize the output contributions of all rules, and the total output is U_total=Σu_k. The defuzzification processing is performed on the aggregated fuzzy output, and the barycenter method, maximum membership degree method or weighted average method is used to convert it into a deterministic control signal. The physical constraints and safety limits of the control signal are considered to ensure that the output signal is within the allowed operating range of the device. The control signal is smoothed to eliminate possible mutations and oscillations in the inference process. The fuzzy control signal contains complete information such as control quantity value, control direction, action time and execution conditions. The fuzzy control signal determined by the execution priority has a clear hierarchical structure and execution logic, and can realize intelligent and accurate control of the carbon capture device.

[0047] The multi-variable regulation channel is established by fuzzy control signals. The fuzzy control signals contain various types of control instructions, which need to be classified and distributed according to the control objects and control targets. According to the control objects, the fuzzy control signals are distributed to the corresponding regulation channels, the temperature control signals are distributed to the heater regulation channel, the pressure control signals are distributed to the valve regulation channel, and the flow control signals are distributed to the pump valve regulation channel. Each regulation channel is responsible for the control of a specific parameter, and the channels realize the collaborative control of multi-variables through the coordination mechanism. The temperature regulation channel controls the temperature of the absorption tower, the regeneration tower and the cooler, the pressure regulation channel controls the system pressure, the steam pressure and the CO2 pressure, and the flow regulation channel controls the amine liquid flow, the steam flow and the cooling water flow. Considering the coupling relationship between the control variables, temperature change will affect the pressure distribution, and flow regulation will affect the mass transfer effect, which need to be decoupled control through feedforward and feedback mechanism. An independent controller and actuator are configured for each regulation channel to ensure accurate execution and rapid response of the control instructions. The multi-variable regulation channel adopts a hierarchical control structure, the upper layer makes coordination decisions, and the lower layer executes specific control. Through the synergistic effect of the regulation channel, the intelligent automatic control of the carbon capture device is realized, which optimizes the purification efficiency and energy consumption level under the premise of ensuring safe operation. The regulation channel records the technical parameters such as control variables, control range, response time and control accuracy, forming a complete multi-variable control system.

[0048] In step S130, the multi-variable regulation channel is subjected to absorption efficiency analysis to generate solvent performance parameters, the lean-rich liquid circulation optimization is implemented based on the solvent performance parameters to form a strengthened absorption domain, and the strengthened absorption domain is subjected to load distribution to generate a parallel capture flow of sub-towers.

[0049] Specifically, the absorption efficiency of the multivariable regulation channel is analyzed to generate solvent performance parameters. The multivariable regulation channel provides synergistic control information of temperature regulation, pressure regulation and flow regulation, which directly affects the absorption performance of the solvent. The efficiency of the CO2 absorption process inside the absorption tower is analyzed, and the absorption efficiency η = (C_in - C_out) / C_in * 100% is calculated, where η is the absorption efficiency, C_in is the inlet CO2 concentration, and C_out is the outlet CO2 concentration. The mass transfer performance of the solvent under different operating conditions is analyzed, and the appropriate operating temperature range of the solvent is determined through the temperature control information of the multivariable regulation channel. When the temperature of the absorption tower is controlled within the appropriate range, the CO2 absorption rate of the MEA solvent reaches a maximum value, and too high a temperature will cause solvent volatilization loss, and too low a temperature will reduce the mass transfer rate. The saturated absorption capacity of the solvent under different pressures is analyzed through the control information of the pressure regulation channel, and an increase in pressure is beneficial to the physical dissolution of CO2, but too high a pressure will increase energy consumption. The chemical stability parameters of the solvent are analyzed, including the solvent degradation rate, corrosive index and thermal stability temperature. The circulation performance indicators of the solvent are calculated, including the absorption load, regeneration efficiency and cycle life. A set of solvent performance parameters is generated, including key data such as absorption capacity, mass transfer coefficient, viscosity characteristics, density change and heat capacity parameters.

[0050] In some embodiments, the lean-rich liquid circulation optimization based on the solvent performance parameters forms a strengthened absorption domain, including: extracting a mass transfer rate index from the solvent performance parameters; adjusting the lean-rich liquid flow ratio to produce an optimal circulation ratio using the mass transfer rate index; applying temperature compensation to the optimal circulation ratio to form adaptive circulation parameters; and constructing a strengthened absorption domain based on the adaptive circulation parameters.

[0051] The mass transfer rate index is extracted from the solvent performance parameters. The relationship between the diffusion coefficient of the solvent and the mass transfer rate is analyzed, and the solvent with a high diffusion coefficient has faster molecular transport capacity. The interfacial tension parameter of the solvent is extracted, and the interfacial tension affects the gas-liquid contact area and the mass transfer effect. The reaction rate constant of the solvent is calculated, and the second-order reaction rate constant k of MEA and CO2 is k = A * exp (-Ea / RT), where A is the pre-exponential factor, Ea is the activation energy, R is the gas constant, and T is the temperature. The influence of the viscosity of the solvent on the mass transfer is analyzed, and low viscosity is beneficial to molecular diffusion but increases the mass transfer resistance. The specific surface area parameter of the solvent is extracted, and the solvent with a large specific surface area provides more mass transfer contact surface. The comprehensive mass transfer rate index MTR = (D * k * A_s) / (μ * σ) is calculated, where MTR is the mass transfer rate index, D is the diffusion coefficient, k is the reaction rate constant, A_s is the specific surface area, μ is the viscosity, and σ is the interfacial tension. The mass transfer rate index under different operating conditions is standardized to eliminate the influence of dimensional differences. The mass transfer rate index reflects the strength of the mass transfer performance of the solvent under certain conditions, and is an important parameter for optimizing the circulation ratio.

[0052] The optimal circulation ratio is determined by adjusting the lean / rich liquid flow ratio using the mass transfer rate indicator. The effect of the lean / rich liquid flow ratio on the mass transfer rate is analyzed. Too large a lean liquid flow rate will result in solvent waste, while too small a flow rate will limit the absorption capacity. A functional relationship between the flow ratio and the mass transfer efficiency is established, η = f(L / G, MTR), where L / G is the liquid-gas ratio and MTR is the mass transfer rate indicator. The optimal flow rate range of the lean / rich liquid is determined through fluid mechanics calculations to ensure a balance between mass transfer effectiveness and energy consumption. The constraints of solvent chemical stability on the circulation ratio are considered. Solvents with high degradation rates need to reduce the circulation intensity to extend the service life, and highly corrosive solvents need to control the circulation flow rate to reduce equipment wear and tear. The optimal circulation ratio R_opt = Q_lean / Q_rich is calculated, where Q_lean is the lean liquid flow rate and Q_rich is the rich liquid flow rate. Through multi-objective optimization, the mass transfer efficiency and energy consumption cost are balanced, and the solvent circulation amount is minimized under the premise of meeting the CO2 removal requirement. The determination of the optimal circulation ratio needs to consider the adaptability of load changes, and the circulation ratio is adjusted under different processing loads to maintain efficient operation.

[0053] Temperature compensation is applied to the optimal circulation ratio to form an adaptive circulation parameter. The optimal circulation ratio is calculated under specific temperature conditions, and temperature changes in actual operation will affect solvent performance and mass transfer effectiveness. The effect of temperature on solvent performance is analyzed. An increase in temperature will reduce the solubility of CO2 in the solvent, but will increase the chemical reaction rate. A temperature compensation function R_comp = R_opt × [1 + α × (T - T_ref)] is established, where R_comp is the compensated circulation ratio, α is the temperature compensation coefficient, T is the actual temperature, and T_ref is the reference temperature. The temperature compensation coefficient is determined according to the thermodynamic properties of the solvent. The effect of seasonal temperature changes on the circulation parameter is considered. The lean liquid circulation amount is increased at high temperatures, and the circulation amount is appropriately reduced at low temperatures. The safety boundary of the compensation parameter is set in combination with the thermal stability temperature of the solvent to ensure that temperature compensation does not cause the solvent to exceed the stable operating range. The adaptive circulation parameter has automatic adjustment capability and can real-time correct the flow ratio according to temperature changes. The adaptive circulation parameter improves the system's adaptability to environmental changes, ensuring stable capture performance under different temperature conditions.

[0054] The enhanced absorption domain is formed by optimizing the solvent circulation parameters and the spatial layout in the absorption tower. The enhanced absorption domain is a region with enhanced mass transfer efficiency, where the solvent is in sufficient contact with CO2, and the mass transfer rate is higher than that in the normal operating state. The spatial position of the lean liquid feed point is determined based on the adaptive circulation parameters. The lean liquid is fed from multiple points at the top of the absorption tower to ensure uniform liquid distribution in the tower. The layout of the rich liquid collection point is optimized, and multiple collection devices are installed at the bottom of the absorption tower to improve the collection efficiency of the rich liquid. The intermediate circulation loop is set up to send part of the rich liquid back to the middle section of the tower for secondary contact to enhance mass transfer. According to the temperature requirements of the adaptive circulation parameters, temperature control devices are installed at different heights to achieve staged control of the solvent temperature. The spatial range and layout characteristics of the enhanced absorption domain provide a capacity basis for subsequent load distribution, and the degree of mass transfer enhancement in the domain directly affects the load carrying capacity. The rationality of the spatial layout is verified by computational fluid dynamics method to ensure uniform flow field distribution and minimum mass transfer resistance. A stable mass transfer environment is formed inside the enhanced absorption domain, the solvent performance is fully utilized, and the CO2 capture efficiency is improved.

[0055] In some embodiments, performing load distribution on the enhanced absorption domain to generate a sub-tower parallel capture flow includes: constructing a capacity evaluation index based on the enhanced absorption domain, the capacity evaluation index including a CO2 absorption rate, a solvent saturation margin, and a tower pressure difference value; implementing dynamic hierarchical load distribution coefficients on the capacity evaluation index; determining a master-slave tower operation mode based on the load distribution coefficients; and generating a sub-tower parallel capture flow via the master-slave tower operation mode.

[0056] A capacity evaluation index is constructed based on the enhanced absorption domain. The CO2 absorption rate index R_abs = dC_CO2 / dt is constructed, where dC_CO2 is the CO2 concentration change amount, and dt is the time interval. This index reflects the CO2 processing capacity of the enhanced absorption domain per unit time. The solvent saturation margin index S_margin = (C_sat - C_current) / C_sat x 100% is constructed, where S_margin is the saturation margin, C_sat is the solvent saturation concentration, and C_current is the current concentration. This index reflects the remaining absorption capacity and load carrying space of the solvent. The tower pressure difference value index ΔP = P_bottom - P_top is constructed, where P_bottom is the tower bottom pressure, and P_top is the tower top pressure. This index reflects the flow resistance and operation stability of the enhanced absorption domain. The three capacity evaluation indexes combined with the spatial layout characteristics of the enhanced absorption domain comprehensively reflect the performance state and load carrying capacity in the domain. The capacity evaluation index is monitored and calculated regularly to form a dynamic capacity evaluation database.

[0057] The capacity evaluation index reflects the real-time performance state of the enhanced absorption domain, and needs to be converted into a quantitative coefficient of load distribution. The CO2 absorption rate is classified, and the high-rate level (R_abs>0.8 mol / (L·s)) corresponds to the distribution coefficient 0.8-1.0, the medium-rate level (0.4-0.8 mol / (L·s)) corresponds to the coefficient 0.5-0.8, and the low-rate level (<0.4 mol / (L·s)) corresponds to the coefficient 0.2-0.5. The solvent saturation margin is classified, and the high-margin level (S_margin>60%) corresponds to the distribution coefficient 0.8-1.0, the medium-margin level (30-60%) corresponds to the coefficient 0.5-0.8, and the low-margin level (<30%) corresponds to the coefficient 0.2-0.5. The inter-tower pressure difference value is classified, and the low-pressure difference level (ΔP<5 kPa) corresponds to the distribution coefficient 0.8-1.0, the medium-pressure difference level (5-15 kPa) corresponds to the coefficient 0.5-0.8, and the high-pressure difference level (>15 kPa) corresponds to the coefficient 0.2-0.5. The comprehensive load distribution coefficient K_load is calculated as w1×K_abs+w2×K_margin+w3×K_pressure, wherein w1, w2, and w3 are weight coefficients, and K_abs, K_margin, and K_pressure are the distribution coefficients of the respective indexes. Considering the correction effect of adaptive circulation parameters on the distribution coefficient, the absorption domain with a higher temperature compensation coefficient appropriately increases its load distribution weight.

[0058] For example, the master-slave tower operation mode is determined based on the load distribution coefficient, which includes: generating a master tower basic load based on the load distribution coefficient for carrying capacity evaluation; deriving a remaining capacity to form a slave tower supplementary load according to the master tower basic load; dynamically matching the master tower basic load and the slave tower supplementary load to generate an inter-tower coordination parameter; and determining the master-slave tower operation mode through the inter-tower coordination parameter integration.

[0059] The main tower basis load is generated by load distribution coefficient based on bearing capacity evaluation. The comprehensive performance state of each absorption tower is reflected by load distribution coefficient K load, and the higher the coefficient, the stronger the bearing capacity. The absorption tower with the highest load distribution coefficient is selected as the candidate main tower, which has the best mass transfer performance and load handling capacity. The theoretical maximum bearing capacity of the main tower C max = K load × Q design is calculated, where C max is the maximum bearing capacity, K load is the load distribution coefficient, and Q design is the design processing capacity. The time-varying characteristics of the load distribution coefficient need to be considered in the bearing capacity evaluation process. When the coefficient fluctuates greatly in a short time, the moving average value is used for smoothing. For the absorption tower with good coefficient stability, the real-time coefficient value can be directly used for calculation. Considering the safety margin and operation stability, the basis load of the main tower is set to a reasonable proportion of the theoretical maximum bearing capacity. The basis load calculation formula is L base = C max × safety factor. The selection of safety factor is determined according to the historical operation stability of the main tower. A higher safety factor is used for the main tower with long-term stable operation, and a lower safety factor is used for the newly operated or frequently malfunctioning main tower. When the load distribution coefficient exceeds the preset threshold, the main tower basis load can be appropriately adjusted; when the coefficient decreases, the basis load is adjusted accordingly to ensure safe operation. The bearing capacity of the main tower changes under different working conditions. The basis load is appropriately reduced during the load peak period, and the basis load can be increased during the load valley period.

[0060] The remaining capacity is formed by the supplementary load of the slave tower according to the main tower basis load. The main tower basis load undertakes the main processing task of the system, and the remaining load needs to be undertaken by the slave tower. The difference between the total processing demand of the system and the main tower basis load is calculated, and the remaining capacity is R remain = Q total - L base, where Q total is the total processing demand of the system. The available processing capacity of the slave tower is analyzed, and its maximum processing capacity is calculated according to the load distribution coefficient of the slave tower. The supplementary load of the slave tower should not exceed the safety range of its maximum processing capacity to avoid overloading of the slave tower. When the remaining capacity exceeds the processing capacity of the slave tower, the third absorption tower needs to be started or the basis load of the main tower needs to be adjusted. The calculation of the supplementary load of the slave tower needs to consider the influence of load fluctuation, and appropriate processing margin is reserved for load peak value. A priority mechanism is established for the supplementary load of the slave tower, and emergency load is allocated to the slave tower first, and regular load is allocated according to the system optimization principle. The supplementary load of the slave tower and the basis load of the main tower form a complementary relationship, and together meet the processing demand of the system.

[0061] The main tower basic load and the from tower supplement load are dynamically proportioned to generate the inter-tower coordination parameters. The main tower basic load and the from tower supplement load need to be proportioned reasonably to realize the coordinated operation of the system. The inter-tower load proportioning R_ratio = L_base / L_supplement is calculated, wherein L_supplement is the from tower supplement load. In the standard operation mode, the load proportioning is controlled within a reasonable range, and the proportioning is too high, which indicates that the from tower is underutilized, and the proportioning is too low, which indicates that the main tower load is too heavy. A dynamic proportioning adjustment mechanism is established, and the load distribution ratio of the main tower and the from tower is adjusted in real time according to the change of the operating condition. A set of inter-tower coordination parameters are generated, including the load distribution ratio, the switching threshold, the response time and the coordination mode and the like control parameters. An inter-tower communication interface is set to realize the information exchange and the coordinated control between the main tower and the from tower. When the main tower load distribution coefficient decreases, the supplement load ratio of the from tower is increased, and when the from tower performance decreases, the basic load ratio of the main tower is increased. The inter-tower coordination parameters ensure the coordinated operation of the main tower and the from tower, and avoid uneven load distribution and system oscillation.

[0062] The main tower and the from tower operation mode are determined through the inter-tower coordination parameter integration. According to the load distribution ratio, the basic operation relationship of the main tower and the from tower is determined, the main tower undertakes the basic load of the system and controls the dominant right, and the from tower undertakes the supplement load and the coordination task. The parameter integration adopts a weighted fusion method, each item of the coordination parameter is allocated a weight according to the importance, and a comprehensive operation mode index is generated. When the load distribution ratio R_ratio > 2.0, the system tends to the main tower dominant mode; when 1.0 < R_ratio < 2.0, the coordinated operation mode is adopted; and when R_ratio < 1.0, the main and from roles are interchanged. When the load distribution coefficient of the main tower is lower than the set threshold, the main and from switching program is started. The switching judgment is based on the comprehensive evaluation of multiple parameters, in addition to the load distribution coefficient, the inter-tower pressure difference, the solvent saturation and the like operation state indexes are also considered. The switching process realizes smooth transition through gradually adjusting the processing load of the two towers, and avoids system disturbance. When a single tower fails, the coordination parameters automatically re-distribute the processing load of the remaining tower body, and the load of the fault tower is distributed according to the bearing capacity proportion of the other towers. The main tower and the from tower operation mode forms multiple sub-modes such as normal coordination, load transfer and emergency operation according to different combinations of the coordination parameters, and adapts to different operating condition requirements.

[0063] The sub-tower parallel capture flow is generated via a master-slave tower operation mode. The sub-tower parallel capture flow refers to a capture operation mode in which multiple absorption towers work cooperatively in a master-slave operation mode, raw gas is processed and purified in parallel after being split, and purified gas is output after being collected. The raw gas is distributed to the master tower and the slave tower according to a load distribution ratio by a splitter, the master tower undertakes the main processing load, and the slave tower undertakes the supplementary processing task, so as to realize reasonable load sharing. The purified gas processed by each tower is collected and output by a collector, so as to ensure the uniformity of the product gas and the stability of the processing effect. A solvent balance pipeline between the towers is arranged, so as to realize solvent allocation and replenishment between the master tower and the slave tower. When the solvent load of the master tower is saturated, fresh solvent can be obtained from the slave tower through the balance pipeline for replenishment. A control system of the sub-tower parallel capture flow is established, so as to uniformly manage the operation parameters and switching operations of each tower, realize coordinated control and state monitoring of the master tower and the slave tower. The sub-tower parallel capture flow adopts modular design, so that the failure or maintenance of a single tower will not affect the normal operation of other towers, and the continuity of the system is ensured. The overall processing capacity and operation reliability of the system are improved through parallel operation. Multiple towers can be started simultaneously during a load peak period, and a single tower can be operated during a load valley period to save energy. The sub-tower parallel capture flow realizes the scaled operation and flexible scheduling of the carbon capture device, and meets the processing requirements under different working conditions.

[0064] In step S140, the sub-tower parallel capture flow is monitored and identified for the degradation of the solvent to obtain the inactivation distribution characteristics, the pollution source is positioned and analyzed based on the inactivation distribution characteristics to generate a solvent updating timing, and the energy consumption of the capture is stepwise adjusted using the solvent updating timing to construct an energy-saving operation configuration.

[0065] Specifically, the solvent degradation monitoring of the parallel capture streams in different towers identifies the deactivation distribution characteristics. The parallel capture streams in different towers provide the solvent circulation information of the collaborative operation of multiple towers, including the solvent state data of the main tower and the slave tower. Online analysis instruments are installed in the solvent circulation pipelines of each absorption tower, including ion chromatograph, ultraviolet spectrometer and conductivity meter, to monitor the chemical composition changes of the solvent in real time. The ion chromatograph detects the degradation products in the MEA solvent, mainly including anion concentrations such as formate, acetate, oxalate and sulfate. The ultraviolet spectrometer monitors the absorbance changes at 280 nm wavelength, which corresponds to the characteristic absorption peak of the MEA degradation products. The conductivity meter measures the ion strength changes of the solvent, and the accumulation of degradation products can significantly increase the conductivity of the solvent. The solvent degradation rate D = (C_degraded) / (C_total) × 100% is calculated, where C_degraded is the degradation product concentration and C_total is the total solvent concentration. The degradation rate distribution at different spatial positions is analyzed, and the degradation rate at the top of the absorption tower is usually lower than 5% due to the lower temperature, and the degradation rate at the bottom of the tower can reach 15-20% due to the higher temperature. The degradation differences between the towers are counted, and the main tower degrades faster than the slave tower due to the heavier load. The deactivation distribution characteristics include key information such as degradation rate value, spatial distribution, time evolution and influencing factors, reflecting the degradation degree and distribution law of the solvent performance.

[0066] In some embodiments, the pollution source positioning analysis based on the deactivation distribution characteristics generates a solvent update timing, including: performing concentration gradient analysis on the deactivation distribution characteristics to identify a degradation area; performing flow path tracking from the degradation area to obtain a pollution source position; prioritizing the pollution source position to generate a processing queue; and generating a solvent update timing based on the processing queue.

[0067] The concentration gradient analysis is performed on the deactivation distribution characteristics to identify the degradation area. The deactivation distribution characteristics show the distribution of the degradation rate at different spatial positions, and the concentration gradient analysis can identify the area with the most dramatic change in degradation rate. The spatial gradient of the degradation rate is calculated as , where is the degradation rate gradient, D is the degradation rate, and x, y, z are spatial coordinates. The gradient size reflects the degree of spatial variation of the degradation rate, and the gradient direction indicates the direction in which the degradation rate increases the fastest. The finite difference method is used to calculate the gradient value of the discrete monitoring points to ensure the accuracy of the gradient calculation. A gradient threshold is set to identify the degradation area, and when the gradient size exceeds 5% / m, it is marked as a degradation area. The degradation area usually appears at positions with large temperature gradients, such as the heater outlet, the heat exchanger and the pipe bend. The geometric characteristics of the degradation area are analyzed, including the area size, volume size and shape characteristics. The severity index of the degradation area is calculated as where S is the severity, and the integral range is the volume of the degradation region. The multiple degradation regions are ranked, and the region with a high severity index is prioritized for treatment. The identification result of the degradation region provides a spatial range limit for subsequent pollution source localization.

[0068] The flow path tracking is performed from the degradation region to obtain the pollution source location. Based on the identified degradation region, the specific location of the pollution source is determined by a fluid mechanics tracking method. The degradation region provides a spatial range of solvent degradation concentration, and the pollution source is usually located upstream or at the center of the degradation region. The particle tracking method is used to simulate the propagation process of the pollutant in the solvent, and the pollution source location is determined by reverse tracking. Virtual tracer particles are released at the boundary position of the degradation region, and the tracking calculation is performed in the reverse direction of the solvent flow. The particle motion equation is dx / dt=-v(x,t), where x is the particle position vector, v is the flow velocity field, and t is the time. The trajectory of the tracked particle is tracked, and the convergence point of the trajectory corresponds to the possible location of the pollution source. The convergence characteristics of different tracking paths are analyzed, and strong convergence indicates that the pollution source location is clear, and weak convergence indicates that the pollution source is widely distributed. Considering the mixing and diffusion effects of the solvent, the influence range of the pollution source will expand over time and distance. The accuracy of the pollution source location is verified by reverse diffusion calculation of the concentration field. When multiple pollution source locations are identified, the interaction and superposition effects are analyzed. The obtained results of the pollution source location include spatial coordinates, influence intensity, diffusion range, and pollution type, etc.

[0069] The pollution source location is prioritized to generate a processing queue. Based on the obtained pollution source location information, the pollution source is sorted and a processing queue is generated by a priority evaluation method. The pollution source location provides the spatial distribution and influence degree of pollution, which needs to be sorted according to the urgency and importance of processing. A pollution source priority evaluation index system is established, including four dimensions of pollution intensity, influence range, diffusion speed and processing difficulty. The pollution intensity is measured by the degradation rate peak value, and the higher the peak value, the higher the priority. The influence range is measured by the spatial volume of pollution diffusion, and the larger the range, the higher the priority. The diffusion speed is measured by the moving speed of the pollution front, and the faster the speed, the higher the priority. The processing difficulty is measured by the technical complexity and cost investment, and the greater the difficulty, the lower the priority. The comprehensive priority index P=w1×I+w2×R+w3×V-w4×D is calculated, where I is the pollution intensity, R is the influence range, V is the diffusion speed, D is the processing difficulty, and w1, w2, w3, w4 are weight coefficients. According to the numerical value of the priority index, the pollution sources are sorted, and the pollution source with the highest index is placed at the head of the queue. Considering the resource limitation and processing capacity, the processing queue is divided into immediate processing, short-term processing and long-term processing three levels. The processing queue contains management information such as pollution source number, priority index, processing time window and resource demand, etc.

[0070] The solvent update schedule is generated based on the processing queue. The solvent update schedule refers to the solvent replacement time plan compiled according to the processing priority of the pollution source. The time node of the solvent update is arranged according to the priority order of the processing queue, and the update task corresponding to the high-priority pollution source is arranged at an earlier time. Considering the requirement of production continuity, the solvent update task is arranged to be performed during a period of low system load or maintenance window. The resource requirements of the solvent update are analyzed, including the amount of fresh solvent, replacement time and manpower input. The optimal batch size of the solvent update is calculated to balance the update efficiency and resource consumption. Time constraints of the solvent update are established to ensure that the update task is completed within the specified time. An emergency update plan is developed to enable a quick response when a sudden serious pollution occurs. The solvent update schedule is represented in the form of a Gantt chart, which directly displays the start time, duration and mutual relationship of each update task. The time sequence planning takes into account the influence of seasonal factors, increases the update frequency during the high-temperature period in summer, and appropriately extends the update interval during the low-temperature period in winter. The solvent update schedule contains complete information such as task number, execution time, update location, solvent amount and quality requirements, providing a systematic execution plan for solvent management.

[0071] An energy-saving operation configuration is constructed by using solvent replacement timing to adjust the energy consumption of capture in stages. The energy-saving operation configuration refers to a time-sharing energy consumption control scheme that matches the energy consumption of the system with the solvent performance state in three stages of low, medium, and high. The solvent replacement timing provides time nodes for the recovery of solvent performance, which correspond to opportunities for changes in system energy consumption. The mechanism of the effect of solvent degradation on system energy consumption is analyzed. The decrease in CO2 absorption capacity of degraded solvent increases the solvent circulation amount, and the accumulation of degradation products increases the regeneration energy consumption. When the concentration of formate in the solvent exceeds 0.1 mol / L, the CO2 loading capacity of the solvent decreases by about 20%, and the circulation flow rate needs to be increased to maintain the capture efficiency. Before solvent replacement, the performance loss is compensated by increasing the circulation flow rate and the regeneration temperature, and at this time, the system energy consumption is in a high-level operation state. For example, when the solvent degradation rate of the main tower reaches 15%, the circulation pump frequency needs to be increased from 50 Hz to 65 Hz, and the regeneration heater power needs to be increased from 800 kW to 1000 kW. After the solvent replacement, the high activity of the fresh solvent allows the system to operate at a lower circulation flow rate and regeneration temperature, achieving a step-down adjustment of energy consumption. Within 48 hours after the completion of solvent replacement, the circulation pump frequency can be gradually reduced to 45 Hz, and the heating power can be reduced to 700 kW, while maintaining the same CO2 capture efficiency. The energy-saving potential ΔE is calculated as ΔE = E_before - E_after, where ΔE is the energy-saving amount (kWh), E_before is the energy consumption before replacement (kWh), and E_after is the energy consumption after replacement (kWh). The step-by-step energy consumption adjustment adopts a three-stage control mode: the first stage is the low-energy consumption operation period after solvent replacement, the second stage is the medium-energy consumption operation period with gradually decreasing solvent performance, and the third stage is the high-energy consumption operation period before solvent replacement. The operation procedures for step-by-step energy consumption adjustment are established, including specific measures such as circulation pump frequency adjustment, regeneration heater power adjustment, and cooler load adjustment. The switching time of energy consumption adjustment is determined based on the monitoring data of solvent performance. When the degradation rate exceeds 10%, it is switched from the first stage to the second stage, and when it exceeds 18%, it is switched to the third stage. The energy-saving operation configuration adopts a time-sharing control strategy, using a low-energy consumption mode when the solvent performance is good, and a high-energy consumption mode when the solvent performance is decreasing. Through the coordination of solvent replacement timing and energy consumption adjustment, the economic and efficient operation of the carbon capture device is realized.

[0072] In step S150, carbon load analysis is performed based on the energy-saving operation configuration to generate a capture capacity curve. Peak-valley scheduling optimization is performed on the capture capacity curve to determine the switching control time. At the switching control time, adaptive adjustment of the working condition is performed to complete the intelligent control of carbon capture.

[0073] Specifically, the carbon load analysis is performed based on the energy-saving operation configuration to generate a capture capacity curve. The time variation of the carbon load is analyzed. The CO2 emission of industrial sources usually has obvious daily and seasonal variation characteristics. The carbon load intensity L(t) = Q CO2(t) x C CO2(t) is calculated in different time periods, where Q CO2(t) is the CO2 gas flow, and C CO2(t) is the CO2 concentration. Combined with the solvent state information in the energy-saving operation configuration, the real-time capture capacity of the system is calculated. In the period after the solvent is updated, the system capture capacity is in a peak state, and the single-tower processing capacity can reach 1200 kg CO2 / h. With the gradual decline of the solvent performance, the capture capacity presents a decay trend, and decreases to 800 kg CO2 / h before the next update. Considering the influence of step energy consumption adjustment, the system capacity is high in the high energy consumption mode, but the operation cost increases, and the system capacity decreases in the low energy consumption mode, but the operation economy is improved. The capture capacity curve is drawn, the horizontal axis is time, and the vertical axis is capture capacity. The curve shape reflects the dynamic variation law of the system processing capacity. The capture capacity curve contains key characteristic parameters such as capacity peak value, valley value, change rate and duration.

[0074] In some embodiments, the peak-valley scheduling optimization is performed on the capture capacity curve to determine the switching control time, including: distinguishing the high load period from the low load period through the capture capacity curve; establishing an operation mode anchor point according to the high load period; performing matching derivation based on the operation mode anchor point to the low load period to determine a mode transition time; and constructing the switching control time by using the mode transition time.

[0075] The high-load period and the low-load period are distinguished by the capture capacity curve. Based on the generated capture capacity curve, the high-load period and the low-load period of system operation are identified by the load analysis method. The capture capacity curve reflects the time variation of the system processing capacity, which can identify the load characteristics combined with the actual carbon load demand. The statistical characteristic parameters of the capture capacity curve are calculated, including the average capacity, the maximum capacity, the minimum capacity, and the standard deviation. Taking the average capacity as the baseline, the period higher than 120% of the average capacity is defined as the high-load period, and the period lower than 80% of the average capacity is defined as the low-load period. The time distribution law of the high-load period and the low-load period is analyzed, and the high-load period usually appears in the peak period of industrial production, such as 9-11 am and 2-4 pm. The low-load period usually appears in the period with less production activities, such as night, weekend, and holidays. The duration and transition frequency of the load period are calculated, and the high-load period lasts for 2-4 hours on average, and the low-load period lasts for 6-8 hours on average. The capacity variation range of the load period is calculated, and the capacity fluctuation range of the high-load period is 1000-1200 kgCO2 / h, and the capacity fluctuation range of the low-load period is 600-800 kgCO2 / h. Considering the transition characteristics between the load periods, the load transition usually needs a transition time of 30-60 minutes. The division results of the high-load period and the low-load period provide a time range for the subsequent selection of the operation mode anchor point.

[0076] The operation mode anchor point is established according to the high-load period. Based on the identified high-load period, the representative running state in the high-load period is selected as the mode anchor point by the mode analysis method. The high-load period represents the high-intensity running state of the system, and its running parameters can be used as the reference benchmark for other periods. The time with the highest capacity and the most stable operation in the high-load period is selected as the main anchor point, and the running parameters of this time represent the optimal performance state of the system. The running parameters of the main anchor point are analyzed, including the solvent circulation flow, the regeneration temperature, the compressor power, and the cooling load. The system state at the anchor point time is recorded, the solvent circulation flow is 180 m³ / h, the regeneration temperature is 120℃, the compressor power is 850 kW, and the capture efficiency is 95%. Auxiliary anchor points are selected in different stages of the high-load period, including the load rising stage anchor point, the stable operation stage anchor point, and the load descending stage anchor point. The parameter differences and variation laws between the anchor points are calculated to quantify the operation mode changes within the high-load period. The anchor point parameters are organized into an operation mode parameter set, and each anchor point corresponds to a complete set of operation parameter configuration. The operation mode anchor point is representative and reproducible, and can guide the system operation under similar working conditions.

[0077] Based on the operating mode anchor point, the matching derivation is implemented in the low load period to determine the mode transition time. Based on the established operating mode anchor point, the mode switching time from high load period to low load period is determined through parameter matching and derivation method. The operating mode anchor point provides the standard operating parameters of the high load period, which need to be adjusted according to the load change. The matching relationship between the anchor point parameters of the high load period and the operating requirements of the low load period is analyzed to identify the parameter items and adjustment range that need to be adjusted. The linear relationship of parameter adjustment is calculated as P_low=P_high×(L_low / L_high)^α, where P_low is the low load period parameter, P_high is the high load period anchor point parameter, L_low is the low load period load, L_high is the high load period load, and α is the adjustment coefficient. The adjustment coefficient is determined according to the response characteristics of different parameters, the adjustment coefficient of solvent circulation flow is 0.8, the adjustment coefficient of regeneration temperature is 0.6, and the adjustment coefficient of compressor power is 1.2. The operating parameter configuration of the low load period is determined through parameter matching and derivation to form a complete operating mode transition scheme. The key nodes of parameter adjustment are identified, and parameter adjustment starts when the load drops to 75% of the anchor point load, and main parameter adjustment is completed when the load drops to 50%. The mode transition time corresponds to the start time and completion time of parameter adjustment, and these times are the nodes where the system operating state changes significantly.

[0078] The switching control time is constructed using the mode transition time. The mode transition time is synchronized with the control cycle of the system to ensure that the switching command is executed within the appropriate control cycle. Considering the response delay and adjustment time of the system, the switching control time is advanced by 30-60 seconds from the mode transition time to reserve sufficient execution time for parameter adjustment. The adjustment priority and time sequence of different parameters are analyzed, and the circulation flow adjustment is prior to temperature adjustment, and the temperature adjustment is prior to pressure adjustment. The time sequence of the switching control time is constructed, and the switching time of each parameter is arranged according to the priority and time sequence. The switching control time contains specific time points, switching parameters, switching amplitudes and execution conditions, etc. The safety check mechanism of switching control is set to check the system state and external conditions before executing the switching to ensure the safety of the switching operation. The feedback mechanism of switching control is established to monitor the effect of switching execution and make necessary compensation adjustment. The switching control time realizes the automatic switching of the system operating mode, automatically adjusts the operating parameters according to the load change, and maintains the efficient and stable operation of the system.

[0079] The working condition self-adaptive adjustment is performed at the switching control moment to complete the intelligent control of carbon capture. Through the determined switching control moment, specific system parameter adjustment and equipment operation switching are performed at the predetermined moment. At the moment of switching from high load to low load, the solvent circulation pump frequency is reduced from 50 Hz to 35 Hz, and the circulation flow rate is reduced from 180 m³ / h to 126 m³ / h. At the same time, the regeneration heater power is reduced from 800 kW to 560 kW, and the regeneration temperature is reduced from 120 DEG C to 100 DEG C. The CO2 compressor operating frequency is reduced from 90% of the rated speed to 70%, and the compression power is reduced by 30% accordingly. At the moment of switching from low load to high load, first, the solvent circulation system of the standby absorption tower is started, and the total circulation flow rate of the system is increased from 126 m³ / h of a single tower to 280 m³ / h of a double tower. The regeneration heater power is increased to 850 kW, and the regeneration temperature is increased from 100 DEG C to 125 DEG C within 15 minutes. According to the carbon load data in the previous 24 hours and the current trend, the parameter adjustment is started 30 minutes in advance to ensure that the system has reached the target operating state when the load peak arrives. By monitoring the pressure change of the solvent circulation pipeline, the CO2 outlet concentration fluctuation and the system energy consumption data, when the CO2 outlet concentration exceeds the set value ± 5 ppm or the system energy consumption deviates from the target value ± 10%, the circulation flow rate and the regeneration temperature are automatically corrected. A three-level parameter adjustment strategy is set: the first level is the circulation flow rate fine adjustment (± 5%), the second level is the temperature adjustment (± 5 DEG C), and the third level is the equipment start-stop switching. When the continuous monitoring of the system abnormal signal exceeds 10 minutes, the automatic switching to the safe operation mode is performed, the processing load is reduced to 60% of the rated capacity, and a maintenance reminder is issued. Through the working condition self-adaptive adjustment, the carbon capture device always maintains a high-efficiency, stable and economic operating state during the whole year operation.

[0080] In order to perform the ship carbon capture intelligent regulation and control method corresponding to the above-mentioned method embodiment, the corresponding functions and technical effects are realized. Referring to Figure 2 , Figure 2 The structure block diagram of the ship carbon capture intelligent regulation and control device 200 provided by the embodiment of the present application is shown. For ease of illustration, only the part related to the present embodiment is shown. The ship carbon capture intelligent regulation and control device 200 provided by the embodiment of the present application comprises:

[0081] The fault diagnosis module 201 is configured to collect the temperature, pressure and flow parameters of the carbon capture device, detect and identify the deviated data segment from the temperature, pressure and flow parameters, perform fault feature analysis on the deviated data segment, and generate a diagnosis mark sequence;

[0082] The fuzzy control module 202 is configured to perform running state reconstruction based on the diagnosis mark sequence to generate a health state flow, implement fuzzy processing on the health state flow to extract a control membership degree, use the control membership degree to stimulate a preset reasoning rule library to generate a fuzzy control signal, and establish a multivariable regulation and control channel through the fuzzy control signal.

[0083] absorption optimization module 203 is configured to perform absorption efficiency analysis on the multivariable regulation channel to generate a solvent performance parameter, implement lean-rich liquid circulation optimization based on the solvent performance parameter to form a strengthened absorption domain, and perform load distribution on the strengthened absorption domain to generate a column-parallel capture stream;

[0084] regeneration management module 204 is configured to perform solvent degradation monitoring on the column-parallel capture stream to identify an inactivation distribution characteristic, perform pollution source positioning analysis based on the inactivation distribution characteristic to generate a solvent update timing, and utilize the solvent update timing to perform step adjustment on capture energy consumption to construct an energy-saving operation configuration;

[0085] intelligent scheduling module 205 is configured to perform carbon load analysis based on the energy-saving operation configuration to generate a capture capacity curve, implement peak-valley scheduling optimization on the capture capacity curve to determine a switching control time, and perform working condition self-adaptive adjustment at the switching control time to complete intelligent control of carbon capture.

[0086] The ship carbon capture intelligent regulation device 200 described above can implement the ship carbon capture intelligent regulation method of the method embodiment described above. The optional items in the method embodiment described above are also applicable to the present embodiment, and will not be described in detail here. The remaining content of the present embodiment can refer to the content of the method embodiment described above, and will not be described in detail in the present embodiment.

[0087] The above embodiments are not an exhaustive enumeration based on the present application, and there can be a plurality of other embodiments not listed. Any replacement and improvement made without violating the concept of the present application is within the protection scope of the present application.

Claims

1. A method for intelligent regulation of carbon capture in a marine vessel, characterized by, The method comprises the following steps: Collecting temperature, pressure and flow parameters of the carbon capture device, detecting abnormal fluctuations of the temperature, pressure and flow parameters to identify deviated data segments, and analyzing fault features of the deviated data segments to generate a diagnostic marker sequence; Based on the diagnostic marker sequence, the running state is reconstructed to generate a health state flow, the health state flow is fuzzified to extract control membership, the control membership is used to trigger a preset reasoning rule library to generate a fuzzy control signal, and a multivariate regulation channel is established through the fuzzy control signal; The absorption efficiency of the multivariate regulation channel is analyzed to generate solvent performance parameters, the solvent performance parameters are used to optimize the lean-liquid and rich-liquid circulation to form a strengthened absorption domain, and the strengthened absorption domain is subjected to load distribution to generate a parallel capture flow of sub-towers; The solvent degradation of the parallel capture flow of sub-towers is monitored to identify the inactivation distribution characteristics, the inactivation distribution characteristics are used for pollution source positioning analysis to generate a solvent update timing, and the solvent update timing is used to perform step-by-step adjustment on the capture energy consumption to build an energy-saving operation configuration; Based on the energy-saving operation configuration, carbon load analysis is performed to generate a capture capacity curve, peak-valley scheduling optimization is performed on the capture capacity curve to determine a switching control time, and working condition self-adaptive adjustment is performed at the switching control time to complete intelligent control of carbon capture.

2. The method of claim 1, wherein, The method comprises the following steps: Extracting abnormal amplitude distribution from the deviated data segments; Applying time correlation to the abnormal amplitude distribution to identify a fault evolution path; Based on the fault evolution path, a failure risk level is constructed; According to the failure risk level, a diagnostic marker sequence is prepared.

3. The method of claim 1, wherein, The method comprises the following steps: From the control membership, an overlapping interval is obtained; According to the overlapping interval, control weight distribution is derived; Based on the control weight distribution, a preset reasoning rule library is triggered to generate an execution priority; Through the execution priority, a fuzzy control signal is determined.

4. The method of claim 1, wherein, The method comprises the following steps: From the solvent performance parameters, a mass transfer rate index is extracted; Using the mass transfer rate index, the lean-liquid and rich-liquid flow ratio is adjusted to generate an optimal circulation ratio; Temperature compensation is applied to the optimal circulation ratio to form an adaptive circulation parameter; Based on the adaptive circulation parameter, a strengthened absorption domain is constructed through spatial layout.

5. The method of claim 1, wherein, The method comprises the following steps: Based on the strengthened absorption domain, a capacity evaluation index is constructed, which includes CO2 absorption rate, solvent saturation margin, and inter-tower pressure difference; The capacity evaluation index is subjected to dynamic grading to obtain a load distribution coefficient; Based on the load distribution coefficient, a master-slave tower operation mode is determined; Through the master-slave tower operation mode, a parallel capture flow of sub-towers is generated.

6. The method of claim 1, wherein, The method comprises the following steps: Concentration gradient analysis is performed on the inactivation distribution characteristics to identify a degradation area; From the degradation area, flow path tracking is performed to obtain the location of the pollution source; The location of the pollution source is prioritized to generate a processing queue; Generate a solvent update timing based on the processing queue.

7. The method of claim 1, wherein, Determine a switching control time point by implementing peak-valley scheduling optimization on the capture capacity curve. Distinguish a high load period from a low load period through the capture capacity curve; Establish an operation mode anchor point according to the high load period; Determine a mode transition time point by implementing matching derivation to the low load period based on the operation mode anchor point; Construct the switching control time point using the mode transition time point.

8. The method of claim 2, wherein, Construct a failure risk level based on the fault evolution path, including: Track a parameter mutation point through the fault evolution path; Establish an association strength between the mutation points and quantify it as a risk transmission coefficient; Start a warning mechanism when the risk transmission coefficient continues to rise, and maintain a monitoring state when the risk transmission coefficient tends to be stable; Form a failure risk level by integrating the trigger frequency of the warning mechanism and the duration of the monitoring state.

9. The method of claim 5, wherein, Determine the master-slave tower operation mode based on the load distribution coefficient, including: Generate a master tower basic load based on the load distribution coefficient for load capacity evaluation; Deduce a remaining capacity to form a slave tower supplementary load according to the master tower basic load; Generate an inter-tower coordination parameter by dynamically proportioning the master tower basic load and the slave tower supplementary load; Determine the master-slave tower operation mode by integrating the inter-tower coordination parameter.

10. A ship carbon capture intelligent regulation device, characterized in that, Including: A fault diagnosis module for collecting temperature, pressure and flow parameters of the carbon capture device, detecting and identifying deviated data segments from the temperature, pressure and flow parameters, and generating a diagnosis marker sequence by analyzing fault characteristics using the deviated data segments; A fuzzy control module for reconstructing a health state stream based on the diagnosis marker sequence to generate a control membership degree by implementing fuzzy processing on the health state stream, and generating a fuzzy control signal by stimulating a pre-set reasoning rule library using the control membership degree, and establishing a multivariate regulation channel through the fuzzy control signal; An absorption optimization module for analyzing the absorption efficiency of the multivariate regulation channel to generate solvent performance parameters, implementing lean-rich liquid circulation optimization based on the solvent performance parameters to form an enhanced absorption domain, and generating a parallel capture stream by distributing loads in the enhanced absorption domain; A regeneration management module for monitoring and identifying inactivation distribution characteristics by degrading the solvent of the parallel capture stream, generating a solvent update timing based on the inactivation distribution characteristics, and constructing an energy-saving operation configuration by step adjusting the capture energy consumption using the solvent update timing; An intelligent scheduling module for generating a capture capacity curve by analyzing carbon load based on the energy-saving operation configuration, determining a switching control time point by implementing peak-valley scheduling optimization on the capture capacity curve, and completing carbon capture intelligent control by performing adaptive adjustment of working conditions at the switching control time point.

Images