Ship scr system intelligent control method, system, device and medium based on digital twinning

Abstract

The application discloses a ship SCR system intelligent control method, system, equipment and medium based on digital twinning, and particularly relates to the technical field of ship SCR intelligent control, wherein through cumulative evaluation and state partition identification of the ammonia injection execution deviation in the time dimension, the ammonia injection execution capability decline trend is identified in advance, and adaptive correction of the control strategy is completed in the digital twinning environment, the SCR emission constraint and the catalyst safe operation window are comprehensively considered in the control process, the generated ammonia injection control strategy is always matched with the actual execution capability and the reaction safety boundary, and the reliability of the control decision under complex working conditions is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for ship SCR (Self-Controlled CR) systems, and more specifically, to intelligent control methods, systems, equipment, and media for ship SCR systems based on digital twins. Background Technology

[0002] During the long-term operation of a ship's SCR system, the ammonia injection system, as a key execution chain connecting control decisions and emission reduction effects, directly determines the actual effectiveness of the SCR control strategy in its physical state. However, in real-world navigation environments, the ammonia injection system is inevitably affected by factors such as high temperature, high humidity, urea crystallization, and frequent start-ups and shutdowns, gradually leading to mechanical degradation phenomena such as nozzle crystallization and valve wear. This results in a nonlinear drift in the mapping relationship between injection volume and control commands, manifesting as injection response hysteresis, saturation, or even intermittent failure. This degradation process is gradual and insidious, making it difficult to observe in real time through direct sensing. The control system often has to rely on indirect back-calculation based on outcome variables such as SCR outlet NOx or ammonia escape, resulting in a hidden deviation where control commands have been issued but actual injection has not met expectations. While existing methods typically assume ideal controllability of the ammonia injection actuator after introducing digital twin technology, assuming that control commands can be accurately executed, this assumption leads to a systematic deviation between the twin model and the physical system when the ammonia injection system experiences nonlinear degradation. This causes the ammonia injection strategy generated based on twin predictions to fail in actual execution or even amplify emission risks. Therefore, the nonlinearity, lack of complete observability, and progressive degradation characteristics of the ammonia injection actuator chain have become key practical technical problems restricting the realization of stable, reliable, and intelligent control of ship SCR systems based on digital twins. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a digital twin-based intelligent control method, system, device and medium for ship SCR systems to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: The intelligent control method for ship SCR system based on digital twin includes the following steps: acquiring the operating data of the ammonia injection system under different main engine loads and exhaust gas conditions, extracting response feature quantities that reflect the injection effectiveness, and obtaining the ammonia injection execution reference feature set; Based on the ammonia injection execution reference feature set, the ammonia injection control command is virtually executed in a twin environment. Combining the predicted nitrogen oxide removal efficiency and ammonia escape trend, the virtual reaction gain after the ammonia injection command is calculated, and the ammonia injection execution deviation descriptor is constructed. Based on the evolution characteristics of the ammonia injection execution deviation descriptor in the time dimension, the injection response hysteresis feature, saturation trend feature and efficiency decay rate feature are extracted to generate the ammonia injection execution state vector. Based on the ammonia injection execution state vector, the operating state of the ammonia injection system is divided into the stable execution region, the degradation transition region and the significant failure region to obtain the execution state classification set. Based on the execution status classification set, the cumulative degree of ammonia injection execution deviation within a continuous control cycle is evaluated in the digital twin. Combined with SCR emission constraints and catalyst safe operation window, it is determined whether there is a risk of execution mismatch in the current ammonia injection control strategy. If so, the ammonia injection control gain, response compensation amount and constraint boundary are adaptively corrected in the twin space.

[0005] In a preferred embodiment, the process of acquiring the operating data of the ammonia injection system under different host loads and exhaust gas conditions, extracting response feature quantities reflecting the injection effectiveness, and obtaining the ammonia injection execution reference feature set is as follows: In the ship's SCR system, control signal acquisition interfaces and emission monitoring sensor units are installed at the ammonia injection control unit, the inlet and outlet of the SCR reactor, respectively, to synchronously collect ammonia injection control command data, SCR inlet nitrogen oxide concentration, SCR outlet nitrogen oxide concentration, ammonia slip concentration and catalytic reaction bed temperature data. The ammonia injection control command data includes ammonia injection valve opening command, injection pulse width and injection frequency; during the acquisition process, the host load, exhaust gas flow rate and exhaust gas temperature information are recorded simultaneously, and the above data are uniformly marked as the original ammonia injection execution operation data; The collected raw ammonia injection execution data is divided into multiple consecutive ammonia injection execution analysis windows according to a preset time window. Within each ammonia injection execution analysis window, the ammonia injection control command data is normalized to obtain the ammonia injection input description vector. Within each ammonia injection analysis window, the nitrogen oxide removal efficiency is calculated based on the nitrogen oxide concentrations at the SCR inlet and outlet, and the ammonia injection reaction output vector is constructed by combining the ammonia escape concentration and the catalytic reaction temperature. In the digital twin space, the ammonia injection input description vector and the ammonia injection reaction output vector are time-aligned to construct a multi-dimensional mapping relationship between ammonia injection commands and reaction results; Based on the multidimensional mapping relationship between ammonia injection command and reaction result, ammonia injection response gain, ammonia injection response hysteresis and ammonia injection efficiency offset are calculated as ammonia injection response characteristic quantities. The ammonia injection response features extracted from each ammonia injection execution analysis window are summarized and labeled and classified according to the host load range and exhaust gas operating condition type to obtain the ammonia injection execution reference feature set.

[0006] In a preferred embodiment, the process of virtually executing ammonia injection control commands in a twin environment based on the ammonia injection execution reference feature set, calculating the virtual reaction gain after ammonia injection command injection by combining the predicted nitrogen oxide removal efficiency and ammonia escape trend, and constructing the ammonia injection execution deviation descriptor is as follows: In a digital twin environment, a virtual execution model of ammonia injection control commands is constructed based on the constructed ammonia injection execution reference feature set. The actual ammonia injection control commands issued during the current control cycle are injected into the virtual execution model to obtain the virtual nitrogen oxide removal efficiency change and virtual ammonia escape change within the corresponding time window, thus obtaining a virtual reaction result set. Calculate the virtual reaction gain of the ammonia injection command based on the virtual reaction result set; During actual system operation, the actual changes in nitrogen oxide removal efficiency and actual ammonia escape within the corresponding time window are acquired synchronously, and the actual ammonia injection reaction gain is calculated in a manner consistent with virtual simulation. The ammonia injection execution deviation is calculated based on the virtual reaction gain and the actual ammonia injection reaction gain, and a descriptive quantity of the ammonia injection execution deviation is constructed.

[0007] In a preferred embodiment, based on the evolution characteristics of the ammonia injection execution deviation descriptor over time, injection response hysteresis features, saturation trend features, and efficiency decay rate features are extracted to generate an ammonia injection execution state vector. The ammonia injection system operating state is then divided into a stable execution region, a degradation transition region, and a significant failure region based on this execution state vector, resulting in the execution state classification set. During the continuous operation of the ammonia injection system, the ammonia injection execution deviation descriptors calculated in each time window are arranged in time sequence according to the control cycle to obtain the ammonia injection execution deviation time series. Based on the ammonia injection execution deviation time series, a sliding time window method is used to perform local trend analysis on the ammonia injection execution deviation descriptor and extract injection response hysteresis features. Based on the extraction of injection response hysteresis characteristics, a nonlinear analysis is performed on the relationship between the ammonia injection execution deviation descriptor and the ammonia injection control command amplitude to extract injection saturation trend characteristics. Regression analysis was performed on the trend of the ammonia injection execution deviation descriptor over a long time scale to extract the ammonia injection efficiency decay rate feature. After obtaining the injection response hysteresis characteristics, injection saturation trend characteristics, and ammonia injection efficiency decay rate characteristics, the above characteristics are combined in a predetermined order to construct the ammonia injection execution state vector. Based on the ammonia injection execution state vector, ammonia injection execution state discrimination rules are pre-established in the digital twin environment, dividing the ammonia injection system operation state into a stable execution zone, a degradation transition zone, and a significant failure zone: when the injection response hysteresis characteristic, injection saturation trend characteristic, and ammonia injection efficiency decay rate characteristic are all within the preset normal threshold range, the ammonia injection system is determined to be in the stable execution zone; when at least one of the above characteristics exceeds the preset warning threshold but does not reach the failure threshold, the ammonia injection system is determined to be in the degradation transition zone; when multiple characteristics exceed the failure threshold simultaneously, the ammonia injection system is determined to be in the significant failure zone, thus forming an ammonia injection execution state classification set.

[0008] In a preferred embodiment, based on the execution state classification set, the cumulative degree of ammonia injection execution deviation within a continuous control cycle is evaluated in the digital twin. Combined with SCR emission constraints and the catalyst safe operating window, it is determined whether the current ammonia injection control strategy has an execution mismatch risk. If so, the process of adaptively correcting the ammonia injection control gain, response compensation, and constraint boundaries in the twin space is as follows: In a digital twin environment, based on the ammonia injection execution status classification set, the operating status of the ammonia injection system is continuously marked according to the control cycle sequence, and the ammonia injection execution deviation description within the same execution status interval is cumulatively evaluated to obtain the ammonia injection execution deviation cumulative sequence. The cumulative sequence of ammonia injection execution deviations is processed by time-weighted integration to calculate the degree of cumulative ammonia injection execution deviations. After obtaining the cumulative degree of ammonia injection execution deviation, it is jointly analyzed with the ammonia injection execution status classification set, and combined with SCR emission constraints and catalyst safe operation window, the conditions for determining the risk of ammonia injection execution mismatch are constructed. Based on the SCR emission constraints and the catalyst safe operating window conditions, calculate the ammonia injection mismatch risk index; When the ammonia injection execution mismatch risk index is less than the preset ammonia injection execution mismatch risk threshold, it is determined that the current ammonia injection control strategy matches the actual execution capability, and the original ammonia injection control strategy remains unchanged. When the ammonia injection execution mismatch risk index is greater than or equal to the preset ammonia injection execution mismatch risk threshold, it is determined that the current ammonia injection control strategy has execution mismatch risk, triggering the ammonia injection control strategy adaptive correction process based on digital twin; After the adaptive correction is triggered, the ammonia injection control parameters are adjusted in the digital twin space.

[0009] In a preferred embodiment, the intelligent control system for a ship's SCR system based on digital twins includes an execution mapping construction module, a virtual execution deviation module, a state evolution partitioning module, and a mismatch assessment and correction module. The execution mapping construction module is used to obtain the operating data of the ammonia injection system under different host loads and exhaust gas conditions, extract the response feature quantities that reflect the injection effectiveness, and obtain the ammonia injection execution reference feature set; The virtual execution deviation module is used to simulate the virtual execution of ammonia injection control commands in a twin environment based on the ammonia injection execution reference feature set. It calculates the virtual reaction gain after the ammonia injection command is injected by combining the predicted nitrogen oxide removal efficiency and ammonia escape change trend, and constructs the ammonia injection execution deviation descriptor. The state evolution partitioning module is used to extract injection response hysteresis features, saturation trend features and efficiency decay rate features based on the evolution characteristics of the ammonia injection execution deviation descriptor in the time dimension, generate an ammonia injection execution state vector, and divide the ammonia injection system operation state into a stable execution zone, a degradation transition zone and a significant failure zone based on the ammonia injection execution state vector, thus obtaining an execution state classification set. The mismatch assessment and correction module is used to assess the cumulative degree of ammonia injection execution deviation in the digital twin based on the execution state classification set. It also combines the SCR emission constraints and the catalyst safe operation window to determine whether there is a risk of execution mismatch in the current ammonia injection control strategy. If so, it adaptively corrects the ammonia injection control gain, response compensation amount and constraint boundary in the twin space.

[0010] The present invention provides a terminal device, including a memory and a processor. The memory is used to store computer-executable instructions, and the processor executes the computer-executable instructions to implement the steps of a digital twin-based intelligent control method for a ship SCR system.

[0011] The present invention provides a computer-readable storage medium storing computer-executable instructions thereon, wherein when the computer-executable instructions are executed by a processor, the steps of a digital twin-based intelligent control method for a ship's SCR system are implemented.

[0012] The technical effects and advantages of this invention are as follows: 1. This invention introduces a reference feature set for ammonia injection execution, a virtual execution gain and an execution deviation evolution modeling mechanism into the digital twin space, making explicit the implicit nonlinear relationship between ammonia injection control commands and actual SCR reaction results, thereby realizing indirect perception and dynamic characterization of the true execution capability of the ammonia injection execution chain, effectively compensating for the systematic deviation caused by the assumption of ideal controllability of the ammonia injection actuator in traditional control methods.

[0013] 2. This invention, through the cumulative evaluation and state zoning of ammonia injection execution deviation over time, can identify the declining trend of ammonia injection execution capability before mechanical degradation such as nozzle crystallization and valve wear causes significant emission exceedances. It then adaptively corrects the control strategy within a digital twin environment, thus avoiding risks such as excessive ammonia injection, ammonia escape amplification, or a sudden drop in denitrification efficiency caused by the continuous superposition of control commands. Simultaneously, it comprehensively considers SCR emission constraints and the catalyst's safe operating window during the control process, ensuring that the generated ammonia injection control strategy always matches the actual execution capability and reaction safety boundary. This improves the robustness and reliability of control decisions under complex operating conditions. Since the entire process does not rely on direct measurement of the physical state of the ammonia injection actuator, it has good versatility and deployability in engineering implementation, significantly improving the emission reduction stability, control reliability, and overall intelligence level of digital twin-based ship SCR systems during long-term operation. Attached Figure Description

[0014] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a flowchart of the method in Embodiment 1 of the present invention; Figure 2 This is a flowchart of the system in Embodiment 2 of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1: Figure 1 The present invention provides an intelligent control method for a ship SCR system based on digital twins, comprising the following steps: The system acquires operating data such as control command sequences, SCR inlet / outlet nitrogen oxide concentrations, ammonia slip, and catalytic reaction temperature under different host loads and exhaust gas conditions. A multidimensional mapping relationship between ammonia injection commands and actual reaction results is constructed in the digital twin space. Response feature quantities reflecting injection effectiveness are extracted to obtain the ammonia injection execution reference feature set. Based on the ammonia injection execution reference feature set, the ammonia injection control command is virtually executed in a twin environment. Combining the predicted nitrogen oxide removal efficiency and ammonia escape trend, the virtual reaction gain after the ammonia injection command is injected is calculated, and an ammonia injection execution deviation descriptor is constructed to characterize the degree of deviation between the control command and the actual injection effect. Based on the evolution characteristics of the ammonia injection execution deviation descriptor in the time dimension, the injection response hysteresis feature, saturation trend feature and efficiency decay rate feature are extracted to generate the ammonia injection execution state vector. Based on the ammonia injection execution state vector, the operating state of the ammonia injection system is divided into the stable execution region, the degradation transition region and the significant failure region to obtain the execution state classification set. Based on the execution state classification set, the cumulative degree of ammonia injection execution deviation within a continuous control cycle is evaluated in the digital twin. Combined with SCR emission constraints and catalyst safe operation window, it is determined whether there is a risk of execution mismatch in the current ammonia injection control strategy. If so, the ammonia injection control gain, response compensation amount and constraint boundary are adaptively modified in the twin space to generate an ammonia injection control strategy that matches the actual execution capability and is then sent to the physical system for execution. This enables the perception, suppression and stable control of nonlinear degradation of the ammonia injection system without relying on direct actuator measurements.

[0017] In this embodiment of the invention, the process of acquiring operational data such as the control command sequence, SCR inlet / outlet nitrogen oxide concentration, ammonia slip, and catalytic reaction temperature of the ammonia injection system under different host loads and exhaust gas conditions, constructing a multidimensional mapping relationship between ammonia injection commands and actual reaction results in a digital twin space, extracting response feature quantities reflecting injection effectiveness, and obtaining the ammonia injection execution reference feature set is as follows: In the ship's SCR system, control signal acquisition interfaces and emission monitoring sensor units are installed at the ammonia injection control unit, the inlet and outlet of the SCR reactor, respectively, to synchronously collect ammonia injection control command data, SCR inlet nitrogen oxide concentration, SCR outlet nitrogen oxide concentration, ammonia slip concentration and catalytic reaction bed temperature data. The ammonia injection control command data includes ammonia injection valve opening command, injection pulse width and injection frequency. Nitrogen oxide concentration data and ammonia escape concentration data are used to characterize the actual reaction results after ammonia injection is executed. During the acquisition process, host load, exhaust gas flow rate and exhaust gas temperature information are recorded simultaneously, and the above data are uniformly marked as the original ammonia injection execution operation data. The collected raw ammonia injection execution data is divided into multiple consecutive ammonia injection execution analysis windows according to a preset time window. It should be noted that the preferred time window length is 2-5 seconds. Understandably, if the time window is too short, the ammonia injection command has not yet fully reacted in the SCR reactor, making it difficult to reflect the true injection effect. If the time window is too long, multiple operating condition change stages may be mixed in, masking the local nonlinear characteristics of the ammonia injection execution. The above-mentioned time window length can cover the main dynamic response process from the ammonia injection command to the SCR reaction result. Within each ammonia injection execution analysis window, the ammonia injection control command data is normalized to obtain the ammonia injection input description vector. It should be noted that the normalization process is as follows: taking the maximum value of the ammonia injection control command within the time window as the benchmark, the ammonia injection valve opening command, injection pulse width and injection frequency are proportionally normalized to form a uniform ammonia injection input description vector. Within each ammonia injection analysis window, the nitrogen oxide removal efficiency is calculated based on the nitrogen oxide concentrations at the SCR inlet and outlet, and the ammonia injection reaction output vector is constructed by combining the ammonia escape concentration and the catalytic reaction temperature. For example, nitrogen oxide removal efficiency The calculation can be performed using the following formula: ,in The volume concentration of nitrogen oxides measured at the SCR inlet. The volume concentration of nitrogen oxides measured at the SCR outlet; It should be noted that the nitrogen oxide removal efficiency is used to reflect the direct contribution of ammonia injection to the emission reduction effect, the ammonia slip concentration is used to reflect the situation of excessive ammonia injection or insufficient mixing, and the catalytic reaction temperature is used to characterize the effectiveness of the ammonia injection reaction conditions. In the digital twin space, the ammonia injection input description vector and the ammonia injection reaction output vector are time-aligned to construct a multi-dimensional mapping relationship between ammonia injection commands and reaction results; Based on the multidimensional mapping relationship between ammonia injection command and reaction result, ammonia injection response gain, ammonia injection response hysteresis and ammonia injection efficiency offset are calculated as ammonia injection response characteristic quantities. For example, in a single time window Inside, among which Using the time window index, calculate the ammonia injection response gain. : ,in for The nitrogen oxide removal efficiency calculated at any given time. for The nitrogen oxide removal efficiency calculated at any given time. for Ammonia injection control input at all times for Ammonia injection control input at all times; The ammonia injection control input is obtained by linearly weighted summation of the normalized ammonia injection valve opening, injection pulse width, and injection frequency. For example, in a single time window Calculate the ammonia injection response hysteresis. : ,in For the maximum allowable physical response delay, The change in ammonia injection control input between adjacent time points. For time offset The change in nitrogen oxide removal efficiency between adjacent time points; For example, in a single time window Calculate the ammonia injection efficiency offset. , ,in This represents the average nitrogen oxide removal efficiency within the time window. For ammonia injection control input With host load The average baseline ammonia injection efficiency; It should be noted that the ammonia injection response gain represents the magnitude of the change in nitrogen oxide removal efficiency caused by the change in ammonia injection command, the ammonia injection response hysteresis represents the time delay between the ammonia injection command and the reaction effect, and the ammonia injection efficiency offset represents the degree of deviation of the actual reaction effect relative to the historical baseline under the same ammonia injection command.

[0018] The ammonia injection response features extracted from each ammonia injection execution analysis window are summarized and labeled and classified according to the host load range and exhaust gas operating condition type to obtain the ammonia injection execution reference feature set. It should be noted that the ammonia injection execution reference feature set is used to describe the actual execution capability and degradation characteristics of the ammonia injection system under different operating conditions. The significance of constructing the ammonia injection execution reference feature set is that, through the multi-dimensional mapping relationship between ammonia injection control commands and actual SCR reaction results, the true execution capability of the ammonia injection execution chain can be indirectly characterized. Thus, without directly measuring nozzle crystallization or valve wear, the nonlinear degradation behavior of the ammonia injection system can be perceived and quantified, providing a reliable basis for intelligent ammonia injection control based on digital twins.

[0019] In this embodiment of the invention, based on the ammonia injection execution reference feature set, the ammonia injection control command is virtually executed and simulated in a twin environment. Combining the predicted nitrogen oxide removal efficiency and ammonia escape trend, the virtual reaction gain after the ammonia injection command is injected is calculated, and an ammonia injection execution deviation descriptor is constructed to characterize the degree of deviation between the control command and the actual injection effect. In a digital twin environment, a virtual execution model of ammonia injection control commands is constructed based on the constructed ammonia injection execution reference feature set, which is used to simulate the SCR reaction behavior corresponding to the ammonia injection commands under ideal execution conditions. It should be noted that the virtual execution model takes the ammonia injection control command vector, host load, and exhaust gas operating parameters as inputs, and outputs the predicted nitrogen oxide removal efficiency, ammonia slip concentration, and catalytic reaction temperature change trends, forming a virtual ammonia injection reaction result set. The virtual execution model can be implemented using any one or a combination of a mechanism-constrained neural network model, a Long Short-Term Memory (LSTM) network model, or a Physical Information Neural Network (PINN) model. The mechanism-constrained neural network model introduces SCR reaction kinetics, mass conservation, and energy conservation constraints into the network structure or loss function to ensure that the model output conforms to real physical laws. The LSTM model is used to characterize the temporal dependence between the ammonia injection command, exhaust gas operating conditions, nitrogen oxide removal efficiency, and ammonia slip, effectively describing the ammonia injection response hysteresis and the influence of historical states on the current reaction. The impact of the results; the physical information neural network model achieves high-precision fitting of complex nonlinear reaction behavior by embedding partial differential equations or empirical reaction equations in the SCR reaction process into the network training process. The above-mentioned virtual execution models are all mature modeling technologies that have been widely used in this field. Their model parameters can be pre-trained based on historical operating data, bench test data or simulation data, and offline convergence verification is completed before deployment to the ship's SCR system. In the actual operation of the system, the above-mentioned virtual execution models are only used as prediction units in the digital twin environment. The pre-trained model parameters are used to quickly perform virtual execution simulation of ammonia injection control commands without the need for complex or high-frequency model retraining in the online operation stage. This ensures prediction accuracy while meeting the engineering application requirements of the ship's SCR system for real-time performance and stability.

[0020] The actual ammonia injection control commands issued during the current control cycle are injected into the virtual execution model to obtain the virtual nitrogen oxide removal efficiency change and virtual ammonia escape change within the corresponding time window, thus obtaining a virtual reaction result set. It should be noted that the virtual nitrogen oxide removal efficiency change is used to characterize the theoretical contribution of the ammonia injection command to the emission reduction effect under ideal execution conditions, and the virtual ammonia slip change is used to characterize the risk of excessive injection introduced by the ammonia injection command under ideal conditions. Calculate the virtual reaction gain of the ammonia injection command based on the virtual reaction result set; For example, virtual reaction gain This describes the overall effect of changes in ammonia injection control commands within a digital twin environment, and its calculation method is as follows: ,in, This represents the change in nitrogen oxide removal efficiency under virtual simulation conditions. This represents the change in ammonia escape concentration under virtual simulation conditions. This indicates the change in the ammonia injection control command. and These are the emission reduction benefit weight and the ammonia escape penalty weight, respectively. It can be understood that by adjusting the weights, the denitrification efficiency and ammonia escape risk can be flexibly balanced under different emission regulations or operating strategies. During actual system operation, the changes in actual nitrogen oxide removal efficiency and actual ammonia slip within the corresponding time window are acquired synchronously, and the actual ammonia injection reaction gain is calculated in a manner consistent with virtual simulation. : ,in, and These represent the actual observed changes in nitrogen oxide removal efficiency and ammonia slip concentration in the physical system, respectively. Calculate the ammonia injection execution deviation based on the virtual reaction gain and the actual ammonia injection reaction gain. Construct a descriptive quantity for ammonia injection execution deviation to quantify the degree of deviation between ammonia injection control commands and actual injection effects; For example, the calculation method for the ammonia injection execution deviation based on the virtual reaction gain and the actual ammonia injection reaction gain is as follows: ; It should be understood that the physical meaning of the ammonia injection execution deviation descriptor is as follows: when the ammonia injection system is in an ideal or slightly degraded state, the virtual reaction gain and the actual reaction gain have a high degree of consistency, and the ammonia injection execution deviation descriptor is small; however, when degradation phenomena such as nozzle crystallization, valve wear, or response hysteresis gradually worsen, the actual reaction gain will be significantly lower than the virtual reaction gain, thus causing the ammonia injection execution deviation descriptor to continuously increase. This descriptor can be used as an important characteristic quantity to characterize the degree of nonlinear degradation of the ammonia injection execution chain.

[0021] In this embodiment of the invention, based on the evolution characteristics of the ammonia injection execution deviation descriptor over time, injection response hysteresis features, saturation trend features, and efficiency decay rate features are extracted to generate an ammonia injection execution state vector. The ammonia injection system operating state is then divided into a stable execution region, a degradation transition region, and a significant failure region based on this execution state vector, resulting in the following process: During the continuous operation of the ammonia injection system, the ammonia injection execution deviation descriptors calculated in each time window are arranged in time sequence according to the control cycle to obtain the ammonia injection execution deviation time series. It should be noted that the ammonia injection execution deviation time series is used to reflect the trend of the deviation between the ammonia injection control command and the actual injection effect over time, and is the basic data for characterizing the evolution of the ammonia injection execution chain's operating state. Based on the ammonia injection execution deviation time series, a sliding time window method is used to perform local trend analysis on the ammonia injection execution deviation descriptor, and the injection response hysteresis feature is extracted. ; It should be noted that the injection response hysteresis characteristic is used to characterize the time delay required for a significant change in the ammonia injection execution deviation after a change in the ammonia injection control command. For example, within a sliding time window, the cross-correlation function between the ammonia injection control command change sequence and the ammonia injection execution deviation sequence is calculated, and the time offset corresponding to the maximum value of the cross-correlation function is used as the injection response hysteresis feature: ,in, These represent correlation calculation functions, such as Pearson correlation coefficient calculation, Spearman's correlation coefficient, and autocorrelation / cross-correlation functions. Indicates the time offset. for Ammonia injection control input at all times Time offset The ammonia injection execution deviation was observed. Based on the extraction of injection response hysteresis characteristics, a nonlinear analysis is performed on the relationship between the ammonia injection execution deviation descriptor and the ammonia injection control command amplitude to extract injection saturation trend characteristics. ; It should be noted that the injection saturation trend feature is used to reflect the trend that when the ammonia injection control command continues to increase, the ammonia injection effect gain gradually decreases or even stops changing. For example, within a preset control command amplitude range, the marginal rate of change of the ammonia injection execution deviation descriptor relative to the change in the ammonia injection control command is calculated, and the degree of decay of the marginal rate of change is used as the injection saturation trend feature: ; Understandably, when the ammonia injection system is in normal condition, the ammonia injection execution deviation changes relatively smoothly with the control command, and the marginal rate of change is small; however, when nozzle crystallization or valve wear limits the injection capacity, the response of the ammonia injection execution deviation to changes in the control command will gradually weaken, showing a clear saturation trend. Regression analysis was performed on the trend of the ammonia injection execution deviation descriptor over a long time scale to extract the ammonia injection efficiency decay rate feature. ; It should be noted that the ammonia injection efficiency decay rate characteristic is used to characterize the rate at which the performance of the ammonia injection system gradually decreases over time. For example, within a preset observation period, a linear fit is performed on the ammonia injection deviation descriptor, and the slope of the fitted curve is used as a feature of the ammonia injection efficiency decay rate: ; After obtaining the injection response hysteresis characteristics, injection saturation trend characteristics, and ammonia injection efficiency decay rate characteristics, these characteristics are combined in a predetermined order to construct the ammonia injection execution state vector: ; It should be noted that the ammonia injection execution state vector is used to comprehensively characterize the dynamic behavior features of the ammonia injection execution chain from multiple dimensions; Based on the ammonia injection execution state vector, ammonia injection execution state discrimination rules are pre-established in the digital twin environment, dividing the ammonia injection system operation state into a stable execution zone, a degradation transition zone, and a significant failure zone: when the injection response hysteresis characteristic, injection saturation trend characteristic, and ammonia injection efficiency decay rate characteristic are all within the preset normal threshold range, the ammonia injection system is determined to be in the stable execution zone; when at least one of the above characteristics exceeds the preset warning threshold but does not reach the failure threshold, the ammonia injection system is determined to be in the degradation transition zone; when multiple characteristics exceed the failure threshold simultaneously, the ammonia injection system is determined to be in the significant failure zone, thus forming an ammonia injection execution state classification set; It should be noted that by extracting temporal features and constructing state vectors based on the ammonia injection execution deviation descriptor, the implicit degradation process of the ammonia injection system can be identified and classified without relying on direct measurement by the ammonia injection actuator.

[0022] In this embodiment of the invention, based on the execution state classification set, the cumulative degree of ammonia injection execution deviation within a continuous control cycle is evaluated in the digital twin. Combined with SCR emission constraints and the catalyst safe operating window, it is determined whether the current ammonia injection control strategy has an execution mismatch risk. If so, the ammonia injection control gain, response compensation, and constraint boundaries are adaptively corrected in the twin space to generate an ammonia injection control strategy that matches the actual execution capability. This strategy is then sent to the physical system for execution. Thus, without relying on direct actuator measurements, the process of sensing, suppressing, and stabilizing the nonlinear degradation of the ammonia injection system is achieved as follows: In a digital twin environment, based on the ammonia injection execution status classification set, the operating status of the ammonia injection system is continuously marked according to the control cycle sequence, and the ammonia injection execution deviation description within the same execution status interval is cumulatively evaluated to obtain the ammonia injection execution deviation cumulative sequence. It should be noted that the cumulative sequence of ammonia injection execution deviation is used to reflect the overall impact of the ammonia injection execution deviation accumulated over time within a continuous control cycle. The cumulative sequence of ammonia injection execution deviations is processed by time-weighted integration to calculate the degree of cumulative ammonia injection execution deviations. For example, the calculation method for the cumulative degree of ammonia injection execution deviation is as follows: ,in, Indicates the first Ammonia injection execution deviation within a historical control cycle This represents the time weight coefficient of the corresponding control cycle, and the closer the control cycle is to the current moment, the greater its weight. It can be understood that by using the time-weighted integral method, the system's sensitivity to the recent trend of ammonia injection degradation can be enhanced, while avoiding excessive interference from long-term historical data on the current judgment. After obtaining the cumulative degree of ammonia injection execution deviation, it is jointly analyzed with the ammonia injection execution status classification set, and combined with SCR emission constraints and catalyst safe operation window, the conditions for determining the risk of ammonia injection execution mismatch are constructed. The SCR emission constraints include nitrogen oxide emission limits and an upper limit for ammonia slip. The safe operating window for the catalyst includes an upper limit for the catalytic reaction temperature, an upper limit for the temperature rise rate, and an allowable range of ammonia injection fluctuations.

[0023] Based on the SCR emission constraints and the catalyst safe operating window conditions, calculate the ammonia injection mismatch risk index; For example, ammonia injection performs mismatch risk indicators The calculation method is as follows: ,in, The preset cumulative threshold for ammonia injection execution deviation. The concentration of nitrogen oxides at the SCR outlet. For emission limits, The catalyst reaction temperature. This refers to the upper limit of the safe temperature for the catalyst. When the ammonia injection execution mismatch risk index is less than the preset ammonia injection execution mismatch risk threshold, it is determined that the current ammonia injection control strategy matches the actual execution capability, and the original ammonia injection control strategy remains unchanged. When the ammonia injection execution mismatch risk index is greater than or equal to the preset ammonia injection execution mismatch risk threshold, it is determined that the current ammonia injection control strategy has execution mismatch risk, triggering the ammonia injection control strategy adaptive correction process based on digital twin; After triggering adaptive correction, the ammonia injection control parameters are adjusted in the digital twin space, including lowering the ammonia injection control gain to avoid excessive command accumulation, dynamically correcting the ammonia injection response compensation amount to compensate for injection hysteresis, and shrinking or shifting the ammonia injection control constraint boundary to ensure that the ammonia injection command is always within the safe operating window of the catalyst. It should be noted that the ammonia injection control gain correction, response compensation adjustment, and constraint boundary correction are all based on the correction rules corresponding to the current ammonia injection execution state interval; For example, the adaptive correction process for the ammonia injection control gain is as follows: In the selected control correction mode, the ammonia injection control gain is dynamically reduced to suppress excessive instruction accumulation caused by decreased execution capability. The correction method is as follows: ,in, For the original ammonia injection control gain, The corrected ammonia injection control gain. The current cumulative deviation in ammonia injection execution. The preset cumulative threshold for ammonia injection execution deviation; For example, the dynamic correction process for the ammonia injection response compensation is as follows: After correcting the ammonia injection control gain, the ammonia injection response compensation amount is adjusted based on the injection response hysteresis characteristics to compensate for the time delay between the ammonia injection command and the actual reaction. The calculation method is as follows: ,in This is the compensation amount for the ammonia injection response. The rate of change of ammonia injection control commands; For example, the adaptive correction process for the ammonia injection control constraint boundary is as follows: After correcting the control gain and response compensation, the ammonia injection control constraint boundary is further narrowed or shifted to ensure that the corrected ammonia injection command always remains within the catalyst's safe operating window. The correction method for the ammonia injection control constraint boundary is as follows: ,in and These are the lower and upper limits of the original ammonia injection control command, respectively. and These are the lower and upper limits of the revised ammonia injection control command, respectively. After completing the adaptive correction of ammonia injection control gain, response compensation amount and constraint boundary, the above correction parameters are injected into the digital twin model to perform virtual execution verification of the corrected ammonia injection control strategy; when the verification results meet the SCR emission constraints and catalyst safe operation conditions, the final ammonia injection control command is generated. It should be noted that, through the above-mentioned hierarchical control parameter adaptive correction mechanism based on the execution state interval, the digital twin system can explicitly introduce the implicit degradation characteristics of the ammonia injection execution chain into the control decision process, and achieve the suppression of nonlinear degradation of the ammonia injection system and the continuous guarantee of control stability without directly sensing the physical state of the ammonia injection actuator. After the ammonia injection control strategy is corrected in the digital twin space, the corrected ammonia injection control strategy is injected into the twin model for virtual verification. After confirming that it meets the emission constraints and catalyst safety constraints within the prediction window, an ammonia injection control strategy that matches the actual ammonia injection execution capability is generated, and the ammonia injection control strategy is sent to the physical SCR system for execution.

[0024] It should be understood that through the above-mentioned adaptive correction process of the control strategy based on the cumulative evaluation of the ammonia injection execution deviation, it is possible to continuously perceive and suppress the nonlinear degradation behavior of the ammonia injection system without directly obtaining the physical state of the ammonia injection actuator, thereby avoiding control failure caused by the decline in ammonia injection execution capability and improving the control stability and reliability of the ship SCR system based on digital twin under complex operating conditions.

[0025] This invention introduces a reference feature set for ammonia injection execution, a virtual execution gain, and an execution deviation evolution modeling mechanism into a digital twin space. This makes explicit the implicit nonlinear relationship between ammonia injection control commands and the actual SCR reaction results, enabling indirect perception and dynamic characterization of the true execution capability of the ammonia injection execution chain. This effectively compensates for the systematic deviations caused by the assumption of ideal controllability of the ammonia injection actuator in traditional control methods.

[0026] This invention, through cumulative evaluation and state zoning of ammonia injection execution deviation over time, can identify the declining trend of ammonia injection execution capability before mechanical degradation such as nozzle crystallization and valve wear causes significant emission exceedances. It then adaptively corrects the control strategy within a digital twin environment, thus avoiding risks such as excessive ammonia injection, ammonia escape amplification, or a sudden drop in denitrification efficiency caused by the continuous superposition of control commands. Simultaneously, it comprehensively considers SCR emission constraints and the catalyst's safe operating window during the control process, ensuring that the generated ammonia injection control strategy always matches the actual execution capability and reaction safety boundary. This improves the robustness and reliability of control decisions under complex operating conditions. Since the entire process does not rely on direct measurement of the physical state of the ammonia injection actuator, it has good versatility and deployability in engineering implementation, significantly improving the emission reduction stability, control reliability, and overall intelligence level of digital twin-based ship SCR systems during long-term operation.

[0027] Example 2: This example introduces an intelligent control system for a ship's SCR system based on digital twins, such as... Figure 2 As shown, it includes an execution mapping construction module, a virtual execution deviation module, a state evolution partitioning module, and a mismatch evaluation and correction module; The execution mapping construction module is used to obtain the operating data of the ammonia injection system under different host loads and exhaust gas conditions, extract the response feature quantities that reflect the injection effectiveness, and obtain the ammonia injection execution reference feature set; The virtual execution deviation module is used to simulate the virtual execution of ammonia injection control commands in a twin environment based on the ammonia injection execution reference feature set. It calculates the virtual reaction gain after the ammonia injection command is injected by combining the predicted nitrogen oxide removal efficiency and ammonia escape change trend, and constructs the ammonia injection execution deviation descriptor. The state evolution partitioning module is used to extract injection response hysteresis features, saturation trend features and efficiency decay rate features based on the evolution characteristics of the ammonia injection execution deviation descriptor in the time dimension, generate an ammonia injection execution state vector, and divide the ammonia injection system operation state into a stable execution zone, a degradation transition zone and a significant failure zone based on the ammonia injection execution state vector, thus obtaining an execution state classification set. The mismatch assessment and correction module is used to assess the cumulative degree of ammonia injection execution deviation within a continuous control cycle in the digital twin based on the execution state classification set. It also combines SCR emission constraints and catalyst safe operation window to determine whether there is a risk of execution mismatch in the current ammonia injection control strategy. If so, it adaptively corrects the ammonia injection control gain, response compensation amount and constraint boundary in the twin space to generate an ammonia injection control strategy that matches the actual execution capability.

[0028] Another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the intelligent control method for a ship SCR system based on digital twins according to any embodiment of the present invention.

[0029] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0030] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and memory. The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0031] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the intelligent control method for a ship SCR system based on digital twin as described in any of the above-described method embodiments of the present invention.

[0032] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0033] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0034] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A digital twin-based intelligent control method for ship SCR systems, characterized by: The process includes the following steps: acquiring the operating data of the ammonia injection system under different host loads and exhaust gas conditions, extracting response feature quantities that reflect the effectiveness of injection, and obtaining the ammonia injection execution reference feature set; Based on the ammonia injection execution reference feature set, the ammonia injection control command is virtually executed in a twin environment. Combining the predicted nitrogen oxide removal efficiency and ammonia escape trend, the virtual reaction gain after the ammonia injection command is calculated, and the ammonia injection execution deviation descriptor is constructed. Based on the evolution characteristics of the ammonia injection execution deviation descriptor in the time dimension, the injection response hysteresis feature, saturation trend feature and efficiency decay rate feature are extracted to generate the ammonia injection execution state vector. Based on the ammonia injection execution state vector, the operating state of the ammonia injection system is divided into the stable execution region, the degradation transition region and the significant failure region to obtain the execution state classification set. Based on the execution status classification set, the cumulative degree of ammonia injection execution deviation within a continuous control cycle is evaluated in the digital twin. Combined with SCR emission constraints and catalyst safe operation window, it is determined whether there is a risk of execution mismatch in the current ammonia injection control strategy. If so, the ammonia injection control gain, response compensation amount and constraint boundary are adaptively corrected in the twin space.

2. The intelligent control method for a ship SCR system based on digital twins according to claim 1, characterized in that: The process of acquiring the operating data of the ammonia injection system under different host loads and exhaust gas conditions, extracting response feature quantities reflecting the injection effectiveness, and obtaining the ammonia injection execution reference feature set is as follows: In the ship's SCR system, control signal acquisition interfaces and emission monitoring sensor units are installed at the ammonia injection control unit, the inlet and outlet of the SCR reactor, respectively, to synchronously collect ammonia injection control command data, SCR inlet nitrogen oxide concentration, SCR outlet nitrogen oxide concentration, ammonia slip concentration and catalytic reaction bed temperature data. The ammonia injection control command data includes ammonia injection valve opening command, injection pulse width and injection frequency; during the acquisition process, the host load, exhaust gas flow rate and exhaust gas temperature information are recorded simultaneously, and the above data are uniformly marked as the original ammonia injection execution operation data; The collected raw ammonia injection execution data is divided into multiple consecutive ammonia injection execution analysis windows according to a preset time window. Within each ammonia injection execution analysis window, the ammonia injection control command data is normalized to obtain the ammonia injection input description vector. Within each ammonia injection analysis window, the nitrogen oxide removal efficiency is calculated based on the nitrogen oxide concentrations at the SCR inlet and outlet, and the ammonia injection reaction output vector is constructed by combining the ammonia escape concentration and the catalytic reaction temperature. In the digital twin space, the ammonia injection input description vector and the ammonia injection reaction output vector are time-aligned to construct a multi-dimensional mapping relationship between ammonia injection commands and reaction results; Based on the multidimensional mapping relationship between ammonia injection command and reaction result, ammonia injection response gain, ammonia injection response hysteresis and ammonia injection efficiency offset are calculated as ammonia injection response characteristic quantities. The ammonia injection response features extracted from each ammonia injection execution analysis window are summarized and labeled and classified according to the host load range and exhaust gas operating condition type to obtain the ammonia injection execution reference feature set.

3. The intelligent control method for a ship SCR system based on digital twins according to claim 2, characterized in that: Based on the ammonia injection execution reference feature set, the ammonia injection control command is virtually executed in a twin environment. Combining the predicted nitrogen oxide removal efficiency and ammonia escape trend, the virtual reaction gain after the ammonia injection command injection is calculated, and the process of constructing the ammonia injection execution deviation descriptor is as follows: In a digital twin environment, a virtual execution model of ammonia injection control commands is constructed based on the constructed ammonia injection execution reference feature set. The actual ammonia injection control commands issued during the current control cycle are injected into the virtual execution model to obtain the virtual nitrogen oxide removal efficiency change and virtual ammonia escape change within the corresponding time window, thus obtaining a virtual reaction result set. Calculate the virtual reaction gain of the ammonia injection command based on the virtual reaction result set; During actual system operation, the actual changes in nitrogen oxide removal efficiency and actual ammonia escape within the corresponding time window are acquired synchronously, and the actual ammonia injection reaction gain is calculated in a manner consistent with virtual simulation. The ammonia injection execution deviation is calculated based on the virtual reaction gain and the actual ammonia injection reaction gain, and a descriptive quantity of the ammonia injection execution deviation is constructed.

4. The intelligent control method for a ship SCR system based on digital twins according to claim 3, characterized in that: Based on the evolution characteristics of the ammonia injection execution deviation descriptor over time, injection response hysteresis features, saturation trend features, and efficiency decay rate features are extracted to generate an ammonia injection execution state vector. Then, based on this vector, the ammonia injection system operating state is divided into a stable execution region, a degradation transition region, and a significant failure region. The process of obtaining the execution state classification set is as follows: During the continuous operation of the ammonia injection system, the ammonia injection execution deviation descriptors calculated in each time window are arranged in time sequence according to the control cycle to obtain the ammonia injection execution deviation time series. Based on the ammonia injection execution deviation time series, a sliding time window method is used to perform local trend analysis on the ammonia injection execution deviation descriptor and extract injection response hysteresis features. Based on the extraction of injection response hysteresis characteristics, a nonlinear analysis is performed on the relationship between the ammonia injection execution deviation descriptor and the ammonia injection control command amplitude to extract injection saturation trend characteristics. Regression analysis was performed on the trend of the ammonia injection execution deviation descriptor over a long time scale to extract the ammonia injection efficiency decay rate feature. After obtaining the injection response hysteresis characteristics, injection saturation trend characteristics, and ammonia injection efficiency decay rate characteristics, the above characteristics are combined in a predetermined order to construct the ammonia injection execution state vector. Based on the ammonia injection execution state vector, ammonia injection execution state discrimination rules are pre-established in the digital twin environment, dividing the ammonia injection system operation state into a stable execution zone, a degradation transition zone, and a significant failure zone: when the injection response hysteresis characteristic, injection saturation trend characteristic, and ammonia injection efficiency decay rate characteristic are all within the preset normal threshold range, the ammonia injection system is determined to be in the stable execution zone; when at least one of the above characteristics exceeds the preset warning threshold but does not reach the failure threshold, the ammonia injection system is determined to be in the degradation transition zone; when multiple characteristics exceed the failure threshold simultaneously, the ammonia injection system is determined to be in the significant failure zone, thus forming an ammonia injection execution state classification set.

5. The intelligent control method for a ship SCR system based on digital twins according to claim 4, characterized in that: Based on the execution state classification set, the cumulative degree of ammonia injection execution deviation within a continuous control cycle is evaluated in the digital twin. Combined with SCR emission constraints and the catalyst safe operating window, it is determined whether the current ammonia injection control strategy has an execution mismatch risk. If so, the process of adaptively correcting the ammonia injection control gain, response compensation, and constraint boundaries in the twin space is as follows: In a digital twin environment, based on the ammonia injection execution status classification set, the operating status of the ammonia injection system is continuously marked according to the control cycle sequence, and the ammonia injection execution deviation description within the same execution status interval is cumulatively evaluated to obtain the ammonia injection execution deviation cumulative sequence. The cumulative sequence of ammonia injection execution deviations is processed by time-weighted integration to calculate the degree of cumulative ammonia injection execution deviations. After obtaining the cumulative degree of ammonia injection execution deviation, it is jointly analyzed with the ammonia injection execution status classification set, and combined with SCR emission constraints and catalyst safe operation window, the conditions for determining the risk of ammonia injection execution mismatch are constructed. Based on the SCR emission constraints and the catalyst safe operating window conditions, calculate the ammonia injection mismatch risk index; When the ammonia injection execution mismatch risk index is less than the preset ammonia injection execution mismatch risk threshold, it is determined that the current ammonia injection control strategy matches the actual execution capability, and the original ammonia injection control strategy remains unchanged. When the ammonia injection execution mismatch risk index is greater than or equal to the preset ammonia injection execution mismatch risk threshold, it is determined that the current ammonia injection control strategy has execution mismatch risk, triggering the ammonia injection control strategy adaptive correction process based on digital twin; After the adaptive correction is triggered, the ammonia injection control parameters are adjusted in the digital twin space.

6. A digital twin-based intelligent control system for ship SCR systems, used to implement the digital twin-based intelligent control method for ship SCR systems as described in any one of claims 1-5, characterized in that: This includes an execution mapping construction module, a virtual execution deviation module, a state evolution partitioning module, and a mismatch evaluation and correction module; The execution mapping construction module is used to obtain the operating data of the ammonia injection system under different host loads and exhaust gas conditions, extract the response feature quantities that reflect the injection effectiveness, and obtain the ammonia injection execution reference feature set; The virtual execution deviation module is used to simulate the virtual execution of ammonia injection control commands in a twin environment based on the ammonia injection execution reference feature set. It calculates the virtual reaction gain after the ammonia injection command is injected by combining the predicted nitrogen oxide removal efficiency and ammonia escape change trend, and constructs the ammonia injection execution deviation descriptor. The state evolution partitioning module is used to extract injection response hysteresis features, saturation trend features and efficiency decay rate features based on the evolution characteristics of the ammonia injection execution deviation descriptor in the time dimension, generate an ammonia injection execution state vector, and divide the ammonia injection system operation state into a stable execution zone, a degradation transition zone and a significant failure zone based on the ammonia injection execution state vector, thus obtaining an execution state classification set. The mismatch assessment and correction module is used to assess the cumulative degree of ammonia injection execution deviation in the digital twin based on the execution state classification set. It also combines the SCR emission constraints and the catalyst safe operation window to determine whether there is a risk of execution mismatch in the current ammonia injection control strategy. If so, it adaptively corrects the ammonia injection control gain, response compensation amount and constraint boundary in the twin space.

7. A terminal device, comprising a memory and a processor, characterized in that... The memory is used to store computer-executable instructions, and when the processor executes the computer-executable instructions, it implements the steps of the intelligent control method for a ship SCR system based on digital twins as described in any one of claims 1-5.

8. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that... When the computer-executable instructions are executed by the processor, they implement the steps of the intelligent control method for a ship SCR system based on digital twins as described in any one of claims 1-5.

Images