A method and system for non-continuous adjustment of ship trim based on hysteresis control

Abstract

This invention discloses a method and system for discontinuous trim adjustment of ships based on hysteresis control. It constructs characteristic quantities of operating conditions as basic indicators for quantifying the magnitude of these changes and sets trigger thresholds as preliminary screening conditions for trend prediction, filtering out scenarios with insignificant operating condition changes. The trend prediction value is used to determine the persistence of operating condition changes, avoiding subsequent judgments triggered by short-term disturbances or instantaneous fluctuations. Accumulated calculations of the duration of sustained stability ensure that the operating condition changes are stable rather than transient. Evaluation of expected energy consumption benefits ensures that trim adjustment has actual energy-saving value. A recovery threshold below the trigger threshold is set to form a hysteresis control range. When the characteristic quantity of the operating condition change falls below the recovery threshold, the discontinuous adjustment cycle is confirmed to have ended, suppressing frequent triggering and recovery switching caused by operating condition fluctuations. This achieves a transition from continuous adjustment to condition-triggered discontinuous adjustment, reducing the workload of the actuators.

Description

Technical Field

[0001] This invention relates to the field of shipbuilding technology, and in particular to a method and system for discontinuous trim adjustment of ships based on hysteresis control. Background Technology

[0002] During navigation, a ship's trim state significantly impacts its drag, propulsion efficiency, and fuel consumption. Proper trim optimization can effectively reduce drag and improve propulsion efficiency, thereby achieving energy conservation and emission reduction. Existing ship trim optimization methods mainly fall into two categories: static and dynamic. Static optimization methods achieve optimal trim before departure by adjusting cargo loading or ballast water distribution, but they struggle to adapt to continuous changes in sea state, speed, and environmental factors during navigation. Dynamic optimization methods adjust trim continuously or frequently based on real-time navigation conditions, but the randomness and volatility of the navigation environment can lead to frequent trim adjustments, increasing the system control burden and causing additional energy consumption due to frequent operation of the ballast water system or related actuators, potentially negating optimization benefits to some extent. Furthermore, continuous control strategies may still trigger adjustments even with small fluctuations in operating conditions, easily causing control system oscillations and affecting operational stability. Therefore, existing technologies generally lack a systematic mechanism for determining the timing of trim adjustments, struggle to effectively reduce adjustment frequency while ensuring adaptability, and are prone to control instability or increased energy consumption in complex navigation environments. Summary of the Invention

[0003] In view of this, the purpose of this invention is to propose a method and system for discontinuous adjustment of ship trim based on hysteresis control.

[0004] To achieve the above-mentioned technical objectives, the technical solution adopted by this invention is as follows: In a first aspect, this application provides a method for discontinuous trim adjustment of a ship based on hysteresis control, comprising: Obtain the ship's current operating status parameters and calculate the characteristic quantities of operating condition changes based on the operating status parameters; When the characteristic quantity of the change in operating condition exceeds the preset trigger threshold, a trend prediction value is generated based on the operating status parameters; Determine whether the trend forecast value indicates the persistence of the change in operating conditions, and calculate or reset the duration of continuous stability based on the determination result; When the duration of sustained stability exceeds the preset stability time threshold, assess the expected energy savings that can be generated by adjusting the ship's trim to different target trim states; If the expected energy consumption benefit exceeds the preset benefit threshold, then the trim adjustment action will be performed to adjust the ship's trim to the target trim state; After the pitch adjustment is executed, the characteristic quantity of the operating condition change is continuously monitored. When the characteristic quantity of the operating condition change falls back to below the preset recovery threshold, it is confirmed that the current discontinuous adjustment cycle has ended and the current pitch state is maintained. The recovery threshold is lower than the trigger threshold. Together, they constitute the hysteresis control range used to suppress frequent adjustments.

[0005] In some embodiments, the current operating status parameters of the ship are obtained, and operating condition change characteristic quantities are calculated based on the operating status parameters, including: Collect data on ship speed, draft, trim angle, main engine power, main engine speed, fuel consumption, and environmental wind, wave, and current to form the original operating status parameters; Perform data preprocessing on the raw operating status parameters, including: The original running state parameters are filled with missing values ​​using an interpolation algorithm to generate the filled running state parameters. Based on the Laida criterion, outlier removal processing is performed on the filled-in running status parameters to generate outlier running status parameters; After removing the data, the speed data, draft data, pitch angle data, main engine power data, main engine speed data, fuel consumption data, and environmental wind, wave and current data from different sampling frequencies in the operating status parameters are uniformly mapped to the same time reference, and time alignment processing is performed to generate aligned operating status parameters. The aligned running state parameters are smoothed by sliding window mean filtering or median filtering to generate smoothed running state parameters. Data standardization processing is performed on the smoothed running state parameters based on Z-score standardization or min-max standardization to generate standardized running state parameters; Standardized operating state parameters corresponding to the ship's previous stable navigation state are extracted from historical navigation data to form a baseline operating condition feature vector. Calculate the Mahalanobis distance between the standardized operating state parameters at the current moment and the baseline operating condition feature vector, and use the calculated Mahalanobis distance value as the characteristic quantity of the operating condition change.

[0006] In some embodiments, when the characteristic quantity of the operating condition change exceeds a preset trigger threshold, a trend prediction value is generated based on the operating status parameters, including: Time series data are extracted from standardized operating status parameters within a preset time window prior to the current moment to form a historical operating status parameter sequence; Time series prediction models are constructed based on historical operating state parameter sequences. These time series prediction models include differential integrated moving average autoregressive models, long short-term memory neural network models, or gated recurrent unit models. The time series prediction model is used to perform prediction processing on standardized operating state parameters within a preset future time window to generate a sequence of future operating state parameters. Perform trend feature extraction processing on the future operating state parameter sequence relative to the standardized operating state parameters at the current time, and output the extracted trend features as the trend prediction value; Among them, the trend characteristics include the linear regression slope value of the future operating state parameter sequence or the cumulative point-by-point difference between the future operating state parameter sequence and the standardized operating state parameter at the current moment.

[0007] In some embodiments, a time series prediction model is used to perform prediction processing on standardized operating state parameters within a preset future time window to generate a future operating state parameter sequence, including: Extract continuous time window data of a preset length from the historical operating state parameter sequence. The continuous time window data contains standardized operating state parameters of the most recent N time points in the historical operating state parameter sequence. Arrange the continuous time window data in chronological order to form the model input feature vector. The model input feature vector is fed into the time series prediction model, and forward inference processing is performed, including: When the time series forecasting model is a differential integrated moving average autoregressive model, the autoregressive moving average calculation is performed; When the time series prediction model is a long short-term memory neural network model or a gated recurrent unit model, the forward propagation calculation of the recurrent neural network is performed step by step to obtain the model output prediction vector. The model output prediction vector contains the predicted values ​​at the standardized scale of each time point within the future preset time window. Obtain the standardization parameters recorded during the data standardization process. The standardization parameters include the mean and standard deviation during Z-score standardization, or the minimum and maximum values ​​during min-max standardization. Based on the standardized parameters, the model output prediction vector is destandardized. The predicted value of each standardized scale in the model output prediction vector is restored to the same physical dimension and numerical scale as the original operating state parameters, generating a sequence of destandardized prediction parameters. Perform validity checks on the destandardized prediction parameter sequence, including: The predicted values ​​at each time point in the destandardized prediction parameter sequence are compared with the ship's physical operating boundary conditions, which include the upper limit of speed, the upper limit of draft, the upper limit of main engine power, and the safe range of the heel angle. If the predicted value at a certain time point in the destandardized prediction parameter sequence exceeds the corresponding ship physical operation boundary conditions, then the predicted value at that time point is corrected to the corresponding boundary value. The predicted parameter sequence, after being destandardized and processed for validity verification, is restandardized and output as the future running state parameter sequence.

[0008] In some embodiments, a trend feature extraction process is performed on the future operating state parameter sequence relative to the standardized operating state parameters at the current time, and the extracted trend features are output as trend prediction values, including: Obtain the future running status parameter sequence, which contains standardized running status parameters at each time point within a future preset time window; Obtain the standardized operating state parameters at the current moment. The standardized operating state parameters at the current moment are the standardized operating state parameters corresponding to the current moment. Perform point-by-point difference calculation on the standardized operating state parameters at each time point in the future operating state parameter sequence and the standardized operating state parameters at the current time to generate a parameter difference sequence. The parameter difference sequence contains the parameter difference at each time point. The parameter difference is the difference obtained by subtracting the standardized operating state parameter at the current time from the standardized operating state parameter at each time point in the future operating state parameter sequence. Linear regression fitting is performed on the parameter difference sequence to obtain linear regression fitting parameters. The linear regression fitting parameters include slope and intercept values. The slope value represents the rate and direction of change of the future running state parameter sequence relative to the current time. Perform cumulative calculation processing on the parameter difference sequence, including: The algebraic sum of the parameter differences at each time point in the parameter difference sequence is calculated to obtain the cumulative difference at each point. The cumulative difference at each point represents the overall change of the parameter sequence in the future operating state relative to the current time. The slope value and the cumulative difference at each point are combined to form a trend feature vector, which contains two dimensions: slope value and cumulative difference at each point. The trend feature vector is output as the trend prediction value.

[0009] In some embodiments, determining whether the trend forecast value indicates the persistence of the change in operating conditions, and calculating or resetting the duration of sustained stability based on the determination result, includes: Obtain the trend forecast value, which includes the slope value and the cumulative point-by-point difference; Obtain preset persistence determination conditions, which include slope value persistence conditions and point-by-point difference accumulation persistence conditions. The slope value persistence condition is that the absolute value of the slope value is greater than the preset slope threshold, and the point-by-point difference accumulation persistence condition is that the absolute value of the point-by-point difference accumulation is greater than the preset accumulation threshold. Compare the slope value in the trend forecast with the slope value persistence condition, and compare the point-by-point difference accumulation in the trend forecast with the point-by-point difference accumulation persistence condition. If the absolute value of the slope value is greater than the preset slope threshold, and the absolute value of the cumulative difference at each point is greater than the preset cumulative threshold, then the trend prediction value indicates that the change in working conditions is continuous. When the trend prediction value indicates that the change in operating conditions is continuous, the start time of the change in operating conditions is obtained. The start time of the change in operating conditions is the time when the characteristic quantity of the change in operating conditions first exceeds the preset trigger threshold. Get the current time, which is the current time point when the continuous stable duration calculation is performed; Calculate the time difference between the current time and the start time of the change in operating conditions, and output the time difference as the duration of continuous stability; Set the operating condition change flag to continuous state. The operating condition change flag is used to mark that the current operating condition is in a continuous state. If the trend forecast value does not meet the persistence determination criteria, the duration of continuous stability will be reset to zero, and the operating condition change flag will be restored to the monitoring state to avoid short-term disturbances from entering the subsequent energy consumption benefit assessment process.

[0010] In some embodiments, when the duration of sustained stability exceeds a preset stability time threshold, the expected energy savings from adjusting the ship's trim to different target trim states are assessed, including: Compare the duration of sustained stability with the stability time threshold; If the duration of sustained stability exceeds the stability time threshold, then obtain the current tilt state and the standardized operating state parameters at the current moment; Based on the standardized operating status parameters at the current moment, the matching candidate target pitch states are queried from the preset target pitch database to generate a set of candidate target pitch states. Obtain the preset energy consumption assessment model, input the current pitch state and the standardized operating state parameters at the current moment into the energy consumption assessment model, perform forward inference calculation, and obtain the predicted energy consumption of the current pitch state; Iterate through each candidate target pitch state in the candidate target pitch state set, input the candidate target pitch state and the standardized operating state parameters at the current moment into the energy consumption assessment model, perform forward inference calculation to obtain the predicted energy consumption of the candidate target pitch state, and calculate the difference between the predicted energy consumption of the current pitch state and the predicted energy consumption of the candidate target pitch state. Use this difference as the candidate expected energy consumption benefit corresponding to the candidate target pitch state. Extract the maximum value from the candidate expected energy consumption benefits corresponding to all candidate target pitch states, determine the candidate target pitch state corresponding to the maximum value as the target pitch state, and output the maximum value as the expected energy consumption benefit.

[0011] In some embodiments, if the expected energy consumption benefit exceeds a preset benefit threshold, a trim adjustment action is performed to adjust the ship's trim to a target trim state, including: If the expected energy consumption benefit is greater than the benefit threshold, then calculate the pitch adjustment between the current pitch state and the target pitch state; Based on the absolute value of the pitch adjustment, the corresponding adjustment rate parameter is retrieved from the preset adjustment rate mapping table to generate a pitch adjustment command containing the target pitch angle value and the adjustment rate parameter. The pitch adjustment command is sent to the pitch adjustment actuator, which then performs the pitch adjustment action. Obtain the actual pitch angle value fed back by the pitch adjustment actuator, and compare the actual pitch angle value with the target pitch state; If the deviation between the actual pitch angle and the target pitch state is greater than or equal to the preset adjustment deviation tolerance threshold, a correction adjustment command is generated and sent to the pitch adjustment actuator again until the deviation is less than the adjustment deviation tolerance threshold.

[0012] In some embodiments, when the characteristic value of the change in operating conditions falls below a preset recovery threshold, confirming the end of the current discontinuous adjustment cycle and maintaining the current pitch state includes: After the tilt adjustment is completed, monitor the characteristic quantities of the operating condition changes; The characteristic quantity of the change in operating condition is compared with the recovery threshold. If the recovery threshold is lower than the preset trigger threshold, the trigger threshold and the recovery threshold together constitute the hysteresis control range. When the characteristic value of the operating condition change falls below the recovery threshold, it is confirmed that the current discontinuous adjustment cycle has ended. Obtain the current pitch state, set the current pitch state as the pitch state corresponding to the new baseline working condition feature, and update the baseline working condition feature vector. Maintain the current tilt state and wait for the next change in operating condition to exceed the trigger threshold.

[0013] In a second aspect, this application provides a ship trim discontinuous adjustment system based on hysteresis control, applicable to the method described in the first aspect. The system includes an operating state parameter acquisition module, an operating condition change characteristic quantity calculation module, a trend prediction module, a continuous stability duration calculation module, an energy consumption benefit assessment module, a trim adjustment execution module, and a hysteresis control monitoring module. The operating state parameter acquisition module is used to acquire the ship's current operating state parameters. The operating condition change characteristic quantity calculation module is connected to the operating state parameter acquisition module and is used to calculate the operating condition change characteristic quantity based on the operating state parameters. The trend prediction module is connected to the operating condition change characteristic quantity calculation module and is used to generate a trend prediction value based on the operating state parameters when the operating condition change characteristic quantity exceeds a preset trigger threshold. The continuous stability duration calculation module is connected to the trend prediction module and is used to determine whether the trend prediction value indicates the continuity of the operating condition change. The system calculates or resets the duration of continuous stability based on the judgment results. The energy consumption benefit assessment module is connected to the continuous stability duration calculation module and is used to assess the expected energy consumption benefit of adjusting the ship's trim to different target trim states when the continuous stability duration exceeds the preset stability time threshold. The trim adjustment execution module is connected to the energy consumption benefit assessment module and is used to execute the trim adjustment action to adjust the ship's trim to the target trim state when the expected energy consumption benefit exceeds the preset benefit threshold. The hysteresis control monitoring module is connected to the trim adjustment execution module and the operating condition change characteristic quantity calculation module and is used to continuously monitor the operating condition change characteristic quantity after the trim adjustment is executed. When the operating condition change characteristic quantity falls below the preset recovery threshold, it is confirmed that the current discontinuous adjustment cycle has ended and the current trim state is maintained. The recovery threshold is lower than the trigger threshold. Together, they constitute the hysteresis control range used to suppress frequent adjustments.

[0014] By adopting the above technical solution, the present invention has the following beneficial effects compared with the prior art: This invention provides a method for discontinuous trim adjustment of a ship based on hysteresis control, comprising: acquiring the ship's current operating state parameters and calculating a characteristic quantity of operating condition change based on the operating state parameters; when the characteristic quantity of operating condition change exceeds a preset trigger threshold, generating a trend prediction value based on the operating state parameters; if the trend prediction value indicates that the operating condition change is persistent, accumulating the duration of sustained stability since the occurrence of the operating condition change; when the duration of sustained stability exceeds a preset stability time threshold, evaluating the expected energy consumption benefits that can be generated by adjusting the ship's trim to different target trim states; if the expected energy consumption benefits exceed a preset benefit threshold, executing a trim adjustment action to adjust the ship's trim to the target trim state; after the trim adjustment is executed, continuously monitoring the characteristic quantity of operating condition change, and when the characteristic quantity of operating condition change falls below a preset recovery threshold, confirming the end of the current discontinuous adjustment cycle and maintaining the current trim state, wherein the recovery threshold is lower than the trigger threshold, and the two together constitute a hysteresis control interval for suppressing frequent adjustments. This invention constructs characteristic quantities of operating conditions as the basic indicators for quantifying the magnitude of operating condition changes, and sets trigger thresholds as preliminary screening conditions for entering trend prediction, thereby filtering out scenarios with insignificant operating condition changes. It uses trend prediction values ​​to determine whether operating condition changes are persistent, avoiding subsequent judgments triggered by short-term disturbances or instantaneous fluctuations. Through the cumulative calculation of continuous stable duration, it ensures that the operating condition changes are stable rather than transient processes. Through the assessment of expected energy consumption benefits, it ensures that tilt adjustment has actual energy-saving value. By setting a recovery threshold below the trigger threshold, a hysteresis control range is formed, and the current discontinuous adjustment cycle is only confirmed to have ended when the characteristic quantity of operating condition changes falls below the recovery threshold, effectively suppressing frequent triggering and recovery switching caused by operating condition fluctuations. Compared with existing technologies, the sequential triggering criterion significantly reduces the frequency of tilt adjustments, avoiding unnecessary adjustment operations; the trend prediction mechanism improves the foresight of adjustment decisions, identifying the difference between short-term disturbances and continuous changes in advance; the energy consumption assessment mechanism avoids ineffective adjustment, preventing the additional energy consumption generated by the adjustment action itself from offsetting the optimization effect; the dynamic hysteresis control mechanism suppresses frequent switching of the system near the threshold, improving the system's operational stability; and it realizes the transformation from continuous adjustment to condition-triggered discontinuous adjustment, reducing the workload of the actuator and improving the feasibility of system engineering. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1This is a schematic diagram of steps S101 to S106 as described in the specific implementation method; Figure 2 This is a schematic diagram of steps S201 to S207 as described in the specific implementation method; Figure 3 This is a schematic diagram of steps S301 to S304 as described in the specific implementation method. Detailed Implementation

[0017] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be particularly noted that the following embodiments are for illustrative purposes only and do not limit the scope of the invention. Similarly, the following embodiments are only some, not all, embodiments of the present invention, and all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 In a first aspect, this embodiment provides a method for discontinuous adjustment of ship trim based on hysteresis control, comprising: S101. Obtain the current operating status parameters of the ship and calculate the characteristic quantities of the operating condition change based on the operating status parameters; S102. When the characteristic quantity of the change in operating condition exceeds the preset trigger threshold, a trend prediction value is generated based on the operating status parameters. S103. Determine whether the trend prediction value indicates the persistence of the change in working conditions, and calculate or reset the duration of continuous stability based on the determination result. S104. When the duration of continuous stability exceeds the preset stability time threshold, assess the expected energy consumption benefits that can be generated by adjusting the ship's trim to different target trim states. S105. If the expected energy consumption benefit exceeds the preset benefit threshold, then perform a trim adjustment action to adjust the ship's trim to the target trim state. S106. After the pitch adjustment is executed, the characteristic quantity of the operating condition change is continuously monitored. When the characteristic quantity of the operating condition change falls back to below the preset recovery threshold, the current discontinuous adjustment cycle is confirmed to have ended and the current pitch state is maintained. The recovery threshold is lower than the trigger threshold. The two together constitute the hysteresis control range used to suppress frequent adjustments.

[0019] In step S101, the operating state parameters refer to the standardized data obtained after preprocessing such as missing value imputation, outlier removal, time alignment, data smoothing, and standardization, collected by shipborne sensors, including ship speed, draft, trim angle, main engine power, main engine speed, fuel consumption, and environmental wind, wave, and current data. The operating condition change characteristic quantity refers to the Mahalanobis distance between the standardized operating state parameters at the current moment and the baseline operating condition characteristic vector at the previous stable navigation state, used to quantify the degree of deviation of the operating condition from the stable state. By constructing quantifiable operating condition change indicators, objective basis is provided for subsequent trigger determination, and multi-dimensional, heterogeneous ship operating data is uniformly mapped to a single characteristic quantity, thereby achieving a numerical representation of operating condition changes.

[0020] In step S102, the trigger threshold is a preset threshold value used to determine whether the change in operating conditions has reached a level requiring attention. Its value can be set through statistical analysis or experience based on ship type, route characteristics, or historical operating data. The trend prediction value refers to the feature quantity extracted after predicting the trend of changes in operating state parameters within a preset future time window using a time series prediction model. This includes the linear regression slope value and the cumulative point-by-point difference, representing the rate and direction of the change in operating conditions, and the overall magnitude, respectively. Trend prediction is only initiated when the characteristic quantity of the change in operating conditions exceeds the trigger threshold, avoiding unnecessary computational resource consumption for minor fluctuations. Simultaneously, the prediction mechanism identifies the persistence characteristics of changes in operating conditions in advance.

[0021] In step S103, the duration of sustained stability refers to the time elapsed from the moment the characteristic quantity of the operating condition change first exceeds the trigger threshold to the current moment. The persistence of the operating condition change is confirmed by a dual threshold determination using both the absolute value of the slope and the absolute value of the accumulated point-by-point difference. The duration is only accumulated after the change is determined to be persistent. This filters out short-term disturbances and instantaneous fluctuations, ensuring that subsequent energy consumption benefit assessments are conducted only when the operating condition change has shown a stable trend, avoiding unnecessary adjustment decisions triggered by transient processes.

[0022] In step S104, the stabilization time threshold is a preset time length threshold used to determine whether the change in operating conditions has stabilized sufficiently. Its value can be determined through actual ship tests or simulation experiments based on the ship's inertia and response characteristics. The expected energy consumption benefit refers to the fuel consumption saved per unit time after adjusting the ship's trim from its current state to a candidate target trim state, calculated using an energy consumption assessment model. By traversing the set of candidate target trim states, the expected energy consumption benefit corresponding to each candidate state is calculated, and the maximum value is extracted as the decision-making basis to ensure that trim adjustment has actual energy-saving value.

[0023] In step S105, the benefit threshold is a preset minimum energy consumption benefit threshold used to determine whether the pitch adjustment has actual energy-saving value. Its setting needs to comprehensively consider the additional energy consumption brought about by the adjustment action itself and the actuator wear. The pitch adjustment command includes the target pitch angle value and the adjustment rate parameter dynamically retrieved from the adjustment rate mapping table based on the absolute value of the pitch adjustment. Through a closed-loop feedback verification mechanism, the actual pitch angle value is obtained after the adjustment is executed and compared with the target value. If the deviation exceeds the adjustment deviation tolerance threshold, a correction command is generated and the adjustment is re-executed until the accuracy requirements are met, ensuring the accuracy and reliability of the adjustment.

[0024] In step S106, the recovery threshold is a preset lower limit value of the operating condition change characteristic quantity below the trigger threshold, which, together with the trigger threshold, constitutes the hysteresis control interval. After the trim adjustment is completed, the operating condition change characteristic quantity is continuously monitored. Only when the characteristic quantity falls back below the recovery threshold is the end of the current discontinuous adjustment cycle confirmed, and the current trim state is updated to the new reference operating condition characteristic, enabling the reference to adaptively follow changes in the ship's operating state. The setting of the hysteresis control interval ensures that when the operating condition change characteristic quantity rises from below or equal to the trigger threshold to above the trigger threshold, the adjustment determination process is triggered. Within the current discontinuous adjustment cycle, the end of the current discontinuous adjustment cycle is confirmed only when the operating condition change characteristic quantity falls back below the recovery threshold. A buffer zone is formed between the trigger threshold and the recovery threshold, effectively suppressing frequent triggering and recovery switching caused by small fluctuations in the operating condition near the trigger threshold.

[0025] This embodiment achieves the transformation of ship trim from continuous adjustment to condition-triggered discontinuous adjustment by constructing a complete closed-loop process: calculation of operating condition change characteristics, trend prediction, determination of continuous stability duration, energy consumption benefit assessment, execution of trim adjustment, and confirmation of hysteresis control. Specifically, the Mahalanobis distance calculation of operating condition change characteristics enables unified quantification of multi-dimensional operating condition changes; the dual threshold determination of trend prediction filters out short-term disturbances; the continuous stability duration ensures the stability of operating condition changes; the energy consumption benefit assessment avoids ineffective adjustment; and the hysteresis control interval suppresses frequent switching. This embodiment significantly reduces the adjustment frequency and additional energy consumption while ensuring trim optimization effects, and improves the stability and engineering feasibility of the control system.

[0026] Please see Figure 2 In some embodiments, the current operating status parameters of the ship are obtained, and operating condition change characteristic quantities are calculated based on the operating status parameters, including: S201. Collect data on ship speed, draft, trim angle, main engine power, main engine speed, fuel consumption, and environmental wind, wave, and current to form the original operating status parameters. S202. Perform data preprocessing on the original operating status parameters, including: The original running state parameters are filled with missing values ​​using an interpolation algorithm to generate the filled running state parameters. Based on the Laida criterion, outlier removal processing is performed on the filled-in running status parameters to generate outlier running status parameters; S203. Map the speed data, draft data, pitch angle data, main engine power data, main engine speed data, fuel consumption data, and environmental wind, wave and current data of different sampling frequencies in the removed operating status parameters to the same time reference, perform time alignment processing, and generate aligned operating status parameters. S204. Perform data smoothing on the aligned running state parameters based on sliding window mean filtering or median filtering to generate smoothed running state parameters. S205. Perform data standardization processing on the smoothed running state parameters based on Z-score standardization or min-max standardization to generate standardized running state parameters; S206. Extract standardized operating state parameters corresponding to the ship's previous stable navigation state from historical navigation data to form a baseline operating condition feature vector. S207. Calculate the Mahalanobis distance between the standardized operating state parameters at the current moment and the baseline operating condition feature vector, and use the calculated Mahalanobis distance value as the operating condition change feature quantity.

[0027] In step S201, the raw operating state parameters refer to the unprocessed ship speed, draft, trim angle, main engine power, main engine speed, fuel consumption, and environmental wind, wave, and current data. These data originate from shipboard sensors, ship energy efficiency monitoring systems, or ship-shore collaborative data platforms. The multi-source data acquisition constructs the foundational data pool for subsequent analysis, and the comprehensiveness of its coverage directly determines the accuracy and robustness of the calculation of operating condition change characteristics.

[0028] In step S202, missing value imputation uses an interpolation algorithm to estimate and fill in missing data using valid data from adjacent time points, avoiding data gaps caused by sensor malfunctions or communication interruptions from affecting subsequent analysis. Outlier removal is based on the Raida criterion, using three standard deviations as the criterion boundary to identify and remove outlier data points deviating from the normal range, eliminating interference introduced by sensor mutations or communication errors. Data cleaning eliminates noise and errors in the original data, ensuring that subsequent processing is based on a reliable data foundation.

[0029] In step S203, time alignment processing maps multi-source sensor data with different sampling frequencies to the same time reference, solving the problem that speed data, draft data, and environmental data cannot be directly analyzed together due to different sampling frequencies. Interpolation or resampling techniques are used to achieve time synchronization of multi-source data, providing time consistency guarantees for subsequent multi-dimensional joint calculations.

[0030] In step S204, data smoothing is performed to reduce noise in the time-aligned data using either sliding window mean filtering or median filtering. Mean filtering is suitable for suppressing random noise, while median filtering is suitable for suppressing impulse noise. This further reduces the interference of sensor measurement noise and short-term random fluctuations on the calculation of operating condition change characteristics, making the extracted operating condition change characteristics more stable and reliable.

[0031] In step S205, data standardization processes map operational state parameters of different dimensions and scales to a unified numerical range using Z-score standardization or min-maximum standardization. Z-score standardization is suitable for scenarios where the data distribution is approximately normal, while min-maximum standardization is suitable for scenarios where the data distribution range is known. This eliminates the weight imbalance problem caused by differences in dimensions of parameters such as speed, draft, and main engine power, ensuring that parameters of each dimension have equal contribution weights in subsequent Mahalanobis distance calculations.

[0032] In step S206, the baseline operating condition feature vector refers to the feature vector composed of standardized operating state parameters corresponding to the ship's previous stable navigation state, which serves as a reference benchmark for measuring changes in operating conditions. By dynamically updating the baseline operating condition feature vector, the system can adaptively follow changes in the ship's long-term operating state, rather than relying on a fixed reference point.

[0033] In step S207, Mahalanobis distance is a distance metric that considers the correlation and dimensional differences of parameters across dimensions, thus eliminating the influence of correlation between different parameters. By calculating the Mahalanobis distance between the standardized operating state parameters at the current moment and the feature vector of the baseline operating condition, multi-dimensional operating condition change information is compressed into a single feature quantity. This feature quantity simultaneously reflects the magnitude of change of each parameter and the change in the correlation structure between parameters, and can more accurately reflect the true degree of change in the operating condition than Euclidean distance.

[0034] This embodiment constructs a complete data processing pipeline from raw data acquisition to the output of operational condition change characteristics, achieving standardized processing of multi-source heterogeneous ship operational data and accurate quantification of operational condition changes. The preprocessing steps, including interpolation imputation, outlier removal, time alignment, data smoothing, and data standardization, progressively improve data quality. Dynamic updates to the baseline operational condition feature vector enable the system to adapt. The introduction of Mahalanobis distance fully considers the correlation between parameters, making the calculation of operational condition change characteristics more scientific and accurate. This provides a reliable data input foundation for subsequent steps such as trend prediction, determination of sustained stability duration, and energy consumption benefit assessment.

[0035] Please see Figure 3 In some embodiments, when the characteristic quantity of the operating condition change exceeds a preset trigger threshold, a trend prediction value is generated based on the operating status parameters, including: S301. Extract time series data from the standardized operating status parameters within the preset time window before the current moment to form a historical operating status parameter sequence; S302. Construct a time series prediction model based on the historical operating status parameter sequence. The time series prediction model includes a differential integrated moving average autoregressive model, a long short-term memory neural network model, or a gated recurrent unit model. S303. Perform prediction processing on standardized operating state parameters within a preset future time window using a time series prediction model to generate a future operating state parameter sequence. S304. Perform trend feature extraction processing on the future operating state parameter sequence relative to the standardized operating state parameters at the current time, and output the extracted trend features as the trend prediction value. Among them, the trend characteristics include the linear regression slope value of the future operating state parameter sequence or the cumulative point-by-point difference between the future operating state parameter sequence and the standardized operating state parameter at the current moment.

[0036] In step S301, the historical operating state parameter sequence refers to a data sequence composed of standardized operating state parameters extracted from a preset time window prior to the current moment, arranged in chronological order. The length of the preset time window can be set according to the ship's dynamic response characteristics and the frequency of changes in the navigation environment. If the window is too long, it may contain outdated historical information; if the window is too short, it may not be able to capture the complete trend of change. By extracting recent historical data as input to the prediction model, it is ensured that the model is based on the data most relevant to the current operating conditions for analysis.

[0037] In step S302, the time series forecasting model is a mathematical model that predicts future data based on historical data. The differentially integrated moving average autoregressive model is suitable for time series data with linear trends and seasonal characteristics. It transforms non-stationary sequences into stationary sequences through differencing operations before forecasting. The long short-term memory neural network model effectively solves the gradient vanishing problem in long-sequence forecasting using traditional recurrent neural networks by introducing forgetting, input, and output gate mechanisms. The gated recurrent unit model, as a simplified variant of the long short-term memory neural network model, reduces computational complexity while maintaining good forecasting performance. By selecting an appropriate time series forecasting model, mathematical support is provided for predicting future operating conditions. The model selection can be adapted according to the characteristics of the actual data and computational resources.

[0038] In step S303, the future operating state parameter sequence refers to the result obtained by predicting the standardized operating state parameters at each time point within a preset future time window using a time series prediction model. The most recent segment of data from the historical operating state parameter sequence is used as the model input feature vector and input into the constructed time series prediction model to perform forward inference calculations, obtaining the predicted values ​​for each future time point. Extrapolating the implicit patterns of change in historical data to the future achieves a forward-looking prediction of the trend of changes in operating conditions.

[0039] In step S304, the trend characteristic refers to the quantitative indicators extracted from the future operating state parameter sequence, used to characterize the persistence and escalation trend of future operating condition changes. The linear regression slope value is obtained by performing linear regression fitting on the future operating state parameter sequence with time points as independent variables and parameter values ​​as dependent variables; its sign and magnitude respectively characterize the direction and rate of future operating condition changes. The cumulative difference at each point is obtained by calculating and summing the differences between the parameter values ​​at each time point in the future operating state parameter sequence and the parameter values ​​at the current time, characterizing the overall magnitude of future operating condition changes. The original prediction results output by the prediction model are transformed into feature quantities with clear physical meaning, providing an intuitive quantitative basis for subsequent persistence determination.

[0040] This embodiment achieves forward-looking prediction of ship operating condition change trends by constructing a complete process of historical sequence extraction, prediction model building, future sequence prediction, and trend feature extraction. The availability of multiple prediction models allows the system to flexibly adapt to the characteristics of actual data; the dual feature extraction method of linear regression slope value and point-by-point difference accumulation characterizes future trends from the two dimensions of change rate and change magnitude, respectively, providing a more comprehensive decision-making basis for subsequent determination of the duration of sustained stability. This elevates the determination of operating condition changes from being based on the current state to being based on future trends, making trim adjustment decisions forward-looking and avoiding unnecessary adjustments triggered by short-term disturbances.

[0041] In some embodiments, a time series prediction model is used to perform prediction processing on standardized operating state parameters within a preset future time window to generate a future operating state parameter sequence, including: Extract continuous time window data of a preset length from the historical operating state parameter sequence. The continuous time window data contains standardized operating state parameters of the most recent N time points in the historical operating state parameter sequence. Arrange the continuous time window data in chronological order to form the model input feature vector. The model input feature vector is fed into the time series prediction model, and forward inference processing is performed, including: When the time series forecasting model is a differential integrated moving average autoregressive model, the autoregressive moving average calculation is performed; When the time series prediction model is a long short-term memory neural network model or a gated recurrent unit model, the forward propagation calculation of the recurrent neural network is performed step by step to obtain the model output prediction vector. The model output prediction vector contains the predicted values ​​at the standardized scale of each time point within the future preset time window. Obtain the standardization parameters recorded during the data standardization process. The standardization parameters include the mean and standard deviation during Z-score standardization, or the minimum and maximum values ​​during min-max standardization. Based on the standardized parameters, the model output prediction vector is destandardized. The predicted value of each standardized scale in the model output prediction vector is restored to the same physical dimension and numerical scale as the original operating state parameters, generating a sequence of destandardized prediction parameters. Perform validity checks on the destandardized prediction parameter sequence, including: The predicted values ​​at each time point in the destandardized prediction parameter sequence are compared with the ship's physical operating boundary conditions, which include the upper limit of speed, the upper limit of draft, the upper limit of main engine power, and the safe range of the heel angle. If the predicted value at a certain time point in the destandardized prediction parameter sequence exceeds the corresponding ship physical operation boundary conditions, then the predicted value at that time point is corrected to the corresponding boundary value. The predicted parameter sequence, after being destandardized and processed for validity verification, is restandardized and output as the future running state parameter sequence.

[0042] In this embodiment, the model input feature vector refers to a continuous time window of data extracted from the historical operating state parameter sequence, containing standardized operating state parameters from the most recent N time points, arranged in chronological order. The selection of the preset length N needs to balance the amount of information in the model input and computational efficiency. If N is too small, it may not be able to capture the complete time dependencies; if N is too large, it may introduce outdated historical information. By constructing a structured model input, data that meets the input format requirements of the time series prediction model is provided.

[0043] In this embodiment, forward inference processing refers to the computation process performed after the model input feature vector is input into the constructed time series prediction model. When the time series prediction model is a differential integrated moving average autoregressive model, an autoregressive moving average calculation is performed, and future values ​​are estimated through a linear combination of the autoregressive term and the moving average term. When the time series prediction model is a long short-term memory neural network model or a gated recurrent unit model, a time-step forward propagation calculation of the recurrent neural network is performed, and the hidden state is updated step-by-step through the gating mechanism and memory unit to output the predicted value. The model output prediction vector contains the predicted values ​​of each time point within a future preset time window at a standardized scale. By extrapolating the implicit change patterns in historical data to the future through the model's forward calculation, numerical prediction of future operating conditions is achieved.

[0044] In this embodiment, the standardized parameters refer to the transformation parameters recorded and saved during the data standardization process in step S205, including the mean and standard deviation during Z-score standardization, or the minimum and maximum values ​​during min-max standardization. Inverse standardization refers to the process of using the standardized parameters to restore the standardized scale prediction values ​​in the model's output prediction vector to the same physical dimensions and numerical scale as the original operating state parameters. Ensuring that the prediction results can be directly compared with physical thresholds (such as revenue thresholds, adjustment bias tolerance thresholds, etc.) in subsequent steps serves as a bridge connecting model prediction and engineering applications.

[0045] In this embodiment, the validity verification process refers to the verification process of comparing the predicted values ​​at each time point in the destandardized prediction parameter sequence with the ship's physical operating boundary conditions. The ship's physical operating boundary conditions include the upper speed limit, upper draft limit, upper main engine power limit, and safe range for heel angle. These boundary conditions are determined by the ship's design parameters and safe operating procedures. When the predicted value at a certain time point exceeds the corresponding boundary condition, the predicted value is corrected to the corresponding boundary value. This ensures the physical rationality of the prediction results and prevents control decisions that exceed the ship's actual operating capabilities due to model prediction errors, which is an important engineering safety assurance measure.

[0046] This embodiment achieves a reliable conversion from historical data to future predictions by constructing a complete process of input feature extraction, forward inference computation, denormalization processing, and validity verification. Specifically, the description of performing differentiated forward inference computation for different types of prediction models reflects a deep understanding of the intrinsic mechanisms of different models; denormalization processing ensures the comparability of prediction results with engineering thresholds; and validity verification introduces physical constraints to ensure that the prediction results do not exceed the actual operational capabilities of the ship. By combining the purely data-driven model prediction results with the ship's physical constraints, the prediction results possess both the accuracy of data-driven approaches and the safety for engineering implementation.

[0047] In some embodiments, a trend feature extraction process is performed on the future operating state parameter sequence relative to the standardized operating state parameters at the current time, and the extracted trend features are output as trend prediction values, including: Obtain the future running status parameter sequence, which contains standardized running status parameters at each time point within a future preset time window; Obtain the standardized operating state parameters at the current moment. The standardized operating state parameters at the current moment are the standardized operating state parameters corresponding to the current moment. Perform point-by-point difference calculation on the standardized operating state parameters at each time point in the future operating state parameter sequence and the standardized operating state parameters at the current time to generate a parameter difference sequence. The parameter difference sequence contains the parameter difference at each time point. The parameter difference is the difference obtained by subtracting the standardized operating state parameter at the current time from the standardized operating state parameter at each time point in the future operating state parameter sequence. Linear regression fitting is performed on the parameter difference sequence to obtain linear regression fitting parameters. The linear regression fitting parameters include slope and intercept values. The slope value represents the rate and direction of change of the future running state parameter sequence relative to the current time. Perform cumulative calculation processing on the parameter difference sequence, including: The algebraic sum of the parameter differences at each time point in the parameter difference sequence is calculated to obtain the cumulative difference at each point. The cumulative difference at each point represents the overall change of the parameter sequence in the future operating state relative to the current time. The slope value and the cumulative difference at each point are combined to form a trend feature vector, which contains two dimensions: slope value and cumulative difference at each point. The trend feature vector is output as the trend prediction value.

[0048] In this embodiment, the future operating state parameter sequence refers to the standardized operating state parameters at each time point within a preset future time window, obtained after validity verification. The standardized operating state parameters at the current moment refer to the standardized operating state parameters corresponding to the current time point, serving as a reference benchmark for measuring future changes.

[0049] In this embodiment, the point-by-point difference calculation process refers to subtracting the parameter value at each time point in the future operating state parameter sequence from the parameter value at the current time point to generate a parameter difference sequence. The difference at each time point in the parameter difference sequence reflects the direction and magnitude of change of each future time point relative to the current time point; positive values ​​indicate that the parameter is increasing, and negative values ​​indicate that the parameter is decreasing.

[0050] In this embodiment, linear regression fitting refers to constructing a linear regression model with time points as independent variables and parameter differences as dependent variables, estimating regression coefficients using the least squares method, and obtaining slope and intercept values. The sign and magnitude of the slope value represent the direction and rate of future operating condition changes, respectively. A positive value indicates that the operating condition change continues to intensify, while a negative value indicates that the operating condition change tends to ease. The larger the absolute value, the faster the rate of change.

[0051] In this embodiment, the cumulative amount calculation process refers to calculating the algebraic sum of the parameter differences at each time point in the parameter difference sequence to obtain the point-by-point cumulative difference amount. The point-by-point cumulative difference amount characterizes the overall magnitude of the change in operating conditions within a future preset time window. The sign of the algebraic sum reflects the overall direction of the change, and the magnitude of the absolute value reflects the overall degree of the change. The slope value and the point-by-point cumulative difference amount jointly characterize the change features of future operating conditions from two dimensions: the rate of change and the magnitude of change. The combination of the two constitutes a change trend feature vector, which is output as a trend prediction value.

[0052] This embodiment transforms the future parameter sequence output by the prediction model into trend characteristic quantities with clear physical meaning, providing an intuitive quantitative basis for subsequent persistence determination. The slope value reflects the direction of the persistence of the trend, while the cumulative difference at each point reflects the cumulative effect of the change. The two complement each other, enabling the trend prediction value to comprehensively characterize the changing features of future operating conditions.

[0053] In some embodiments, determining whether the trend forecast value indicates the persistence of the change in operating conditions, and calculating or resetting the duration of sustained stability based on the determination result, includes: Obtain the trend forecast value, which includes the slope value and the cumulative point-by-point difference; Obtain preset persistence determination conditions, which include slope value persistence conditions and point-by-point difference accumulation persistence conditions. The slope value persistence condition is that the absolute value of the slope value is greater than the preset slope threshold, and the point-by-point difference accumulation persistence condition is that the absolute value of the point-by-point difference accumulation is greater than the preset accumulation threshold. Compare the slope value in the trend forecast with the slope value persistence condition, and compare the point-by-point difference accumulation in the trend forecast with the point-by-point difference accumulation persistence condition. If the absolute value of the slope value is greater than the preset slope threshold, and the absolute value of the cumulative difference at each point is greater than the preset cumulative threshold, then the trend prediction value indicates that the change in working conditions is continuous. When the trend prediction value indicates that the change in operating conditions is continuous, the start time of the change in operating conditions is obtained. The start time of the change in operating conditions is the time when the characteristic quantity of the change in operating conditions first exceeds the preset trigger threshold. Get the current time, which is the current time point when the continuous stable duration calculation is performed; Calculate the time difference between the current time and the start time of the change in operating conditions, and output the time difference as the duration of continuous stability; The operating condition change flag is set to continuous. The operating condition change flag is used to mark the current state as continuous operating condition change.

[0054] If the trend forecast value does not meet the persistence determination criteria, the duration of continuous stability will be reset to zero, and the operating condition change flag will be restored to the monitoring state to avoid short-term disturbances from entering the subsequent energy consumption benefit assessment process.

[0055] In this embodiment, the trend prediction value includes two dimensions: slope value and cumulative difference at each point, which respectively characterize the rate and direction of future changes in operating conditions and the overall magnitude. The preset slope threshold and preset cumulative difference threshold are pre-set threshold values ​​used to determine whether changes in operating conditions are persistent. Their values ​​can be determined through statistical analysis based on ship type, route characteristics, or historical operating data.

[0056] In this embodiment, persistence is determined through two conditions: the absolute value of the slope is greater than a preset slope threshold, indicating that the change in operating conditions has a clear direction and a sufficient rate; the absolute value of the cumulative difference at each point is greater than a preset cumulative threshold, indicating that the change in operating conditions has a sufficient overall magnitude. Only when both conditions are met is the change in operating conditions considered persistent. This dual-determination mechanism is more robust than a single-indicator determination and can effectively filter out false trends and short-term fluctuations.

[0057] In this embodiment, the start time of the operating condition change refers to the moment when the characteristic quantity of the operating condition change first exceeds the preset trigger threshold, serving as the starting reference point for calculating the continuous stable duration. The continuous stable duration refers to the time difference between the current moment and the start time of the operating condition change, representing the length of time the operating condition change has lasted. The accumulation of this duration begins when the characteristic quantity of the operating condition change first exceeds the trigger threshold, ensuring that the starting point of the duration strictly corresponds to the trigger point of the operating condition change.

[0058] This embodiment achieves accurate determination of the persistence of operating condition changes and accurate accumulation of duration through a dual threshold determination mechanism and start time tracing. The dual determination of slope value and point-by-point difference accumulation comprehensively evaluates both the rate of change and the magnitude of change, avoiding misjudgments that might occur with a single indicator. Accurate tracing of the start time of operating condition changes provides a reliable duration basis for subsequent comparison of stable time thresholds.

[0059] In some embodiments, when the duration of sustained stability exceeds a preset stability time threshold, the expected energy savings from adjusting the ship's trim to different target trim states are assessed, including: Compare the duration of sustained stability with the stability time threshold; If the duration of sustained stability exceeds the stability time threshold, then obtain the current tilt state and the standardized operating state parameters at the current moment; Based on the standardized operating status parameters at the current moment, the matching candidate target pitch states are queried from the preset target pitch database to generate a set of candidate target pitch states. Obtain the preset energy consumption assessment model, input the current pitch state and the standardized operating state parameters at the current moment into the energy consumption assessment model, perform forward inference calculation, and obtain the predicted energy consumption of the current pitch state; Iterate through each candidate target pitch state in the candidate target pitch state set, input the candidate target pitch state and the standardized operating state parameters at the current moment into the energy consumption assessment model, perform forward inference calculation to obtain the predicted energy consumption of the candidate target pitch state, and calculate the difference between the predicted energy consumption of the current pitch state and the predicted energy consumption of the candidate target pitch state. Use this difference as the candidate expected energy consumption benefit corresponding to the candidate target pitch state. Extract the maximum value from the candidate expected energy consumption benefits corresponding to all candidate target pitch states, determine the candidate target pitch state corresponding to the maximum value as the target pitch state, and output the maximum value as the expected energy consumption benefit.

[0060] In this embodiment, the stabilization time threshold is a preset time length threshold used to determine whether the change in operating conditions has stabilized sufficiently. Its value can be determined through actual ship tests or simulation experiments based on the ship's inertia and response characteristics. The current trim state refers to the actual trim angle of the ship at the current moment, serving as the benchmark state for energy consumption assessment.

[0061] In this embodiment, the candidate target pitch state set is a collection of multiple candidate target pitch angle values ​​obtained by querying and matching standardized operating state parameters from a preset target pitch database at the current moment. The target pitch database stores the optimal pitch angle values ​​corresponding to different operating conditions, which can be obtained through historical navigation data statistics or simulation optimization. This ensures that the candidate target pitch states match the current operating conditions, rather than being fixed or randomly selected.

[0062] In this embodiment, the energy consumption assessment model is a regression model that takes the pitch angle and operating status parameters as inputs and fuel consumption per unit time as output. It can be trained based on historical navigation data using machine learning or statistical regression methods. The predicted energy consumption under the current pitch condition refers to the predicted value calculated after inputting the current pitch condition and current operating status parameters into the model, which serves as the benchmark value for energy consumption comparison.

[0063] In this embodiment, the calculation process iterates through each candidate state in the set of candidate target pitch states, inputting it along with the current operating state parameters into the energy consumption assessment model to calculate the predicted energy consumption under that candidate state. The difference between this predicted energy consumption and the current pitch state's predicted energy consumption is then calculated to obtain the candidate expected energy consumption benefit corresponding to that candidate state. The maximum value is extracted from all candidate expected energy consumption benefits, and the candidate target pitch state corresponding to this maximum value is determined as the target pitch state. This maximum value is then output as the expected energy consumption benefit.

[0064] This embodiment achieves an objective assessment of the energy-saving benefits of tilt adjustment through quantitative calculations using an energy consumption assessment model and traversal comparisons of candidate states. Specifically, matching the candidate target tilt state with the current operating condition ensures the assessment's relevance, the data-driven nature of the energy consumption assessment model guarantees its accuracy, and the maximum value extraction strategy ensures that adjustment decisions are guided by maximizing energy-saving benefits.

[0065] In some embodiments, if the expected energy consumption benefit exceeds a preset benefit threshold, a trim adjustment action is performed to adjust the ship's trim to a target trim state, including: If the expected energy consumption benefit is greater than the benefit threshold, then calculate the pitch adjustment between the current pitch state and the target pitch state; Based on the absolute value of the pitch adjustment, the corresponding adjustment rate parameter is retrieved from the preset adjustment rate mapping table to generate a pitch adjustment command containing the target pitch angle value and the adjustment rate parameter. The pitch adjustment command is sent to the pitch adjustment actuator, which then performs the pitch adjustment action. Obtain the actual pitch angle value fed back by the pitch adjustment actuator, and compare the actual pitch angle value with the target pitch state; If the deviation between the actual pitch angle and the target pitch state is greater than or equal to the preset adjustment deviation tolerance threshold, a correction adjustment command is generated and sent to the pitch adjustment actuator again until the deviation is less than the adjustment deviation tolerance threshold.

[0066] In this embodiment, the benefit threshold is a preset minimum energy consumption benefit threshold used to determine whether pitch adjustment has actual energy-saving value. Its setting needs to comprehensively consider the additional energy consumption brought about by the adjustment action itself and the wear and tear on the actuator, to avoid ineffective adjustment due to insufficient benefit to cover adjustment costs. The pitch adjustment amount refers to the difference between the target pitch state and the current pitch state; its sign and magnitude represent the direction and amplitude of the adjustment, respectively.

[0067] In this embodiment, the adjustment rate mapping table is a preset correspondence table between the absolute value of the pitch adjustment and the adjustment rate parameter. The adjustment rate parameter is dynamically determined according to the magnitude of the adjustment. A higher rate is used for large adjustments to shorten the adjustment time, and a lower rate is used for small adjustments to avoid excessive impact. The pitch adjustment command is a control command containing the target pitch angle value and the adjustment rate parameter, which is sent to the pitch adjustment actuator for execution.

[0068] In this embodiment, the pitch adjustment actuator includes a ballast water system or an attitude adjustment device, which changes the pitch angle by adjusting the ballast water distribution or the attitude control device. The adjustment deviation tolerance threshold is a preset maximum allowable deviation value used to determine whether the adjustment accuracy meets the requirements. When the deviation between the actual pitch angle value and the target pitch state is greater than or equal to this threshold, a correction adjustment command is generated and sent back to the actuator. Through a closed-loop feedback mechanism, the value is gradually approached until the deviation meets the accuracy requirements.

[0069] This embodiment achieves precise execution of tilt adjustment through dynamic determination of the adjustment rate and a closed-loop feedback verification mechanism. The adjustment rate mapping table adaptively matches the adjustment speed and amplitude, balancing adjustment efficiency and execution stability. The closed-loop feedback verification mechanism ensures that the adjustment accuracy meets engineering requirements through deviation detection and iterative execution of correction commands, avoiding the accumulation of errors that may occur during a single adjustment.

[0070] In some embodiments, when the characteristic value of the change in operating conditions falls below a preset recovery threshold, confirming the end of the current discontinuous adjustment cycle and maintaining the current pitch state includes: After the tilt adjustment is completed, monitor the characteristic quantities of the operating condition changes; The characteristic quantity of the change in operating condition is compared with the recovery threshold. If the recovery threshold is lower than the preset trigger threshold, the trigger threshold and the recovery threshold together constitute the hysteresis control range. When the characteristic value of the operating condition change falls below the recovery threshold, it is confirmed that the current discontinuous adjustment cycle has ended. Obtain the current pitch state, set the current pitch state as the pitch state corresponding to the new baseline working condition feature, and update the baseline working condition feature vector. Maintain the current tilt state and wait for the next change in operating condition to exceed the trigger threshold.

[0071] In this embodiment, the recovery threshold is a preset lower limit value of the operating condition change characteristic quantity below the trigger threshold. The trigger threshold and the recovery threshold together constitute the hysteresis control interval. The setting of the hysteresis control interval ensures that the operating condition change characteristic quantity needs to rise from a state below or equal to the trigger threshold to above the trigger threshold to trigger the adjustment process, and then fall back below the recovery threshold to confirm the end of this discontinuous adjustment cycle. A buffer zone is formed between triggering and recovery. This effectively suppresses frequent triggering and recovery switching caused by small fluctuations in the operating condition change characteristic quantity near the trigger threshold, improving the stability of the control system.

[0072] In this embodiment, the reference operating condition feature vector refers to the feature vector composed of standardized operating state parameters corresponding to the ship's previous stable navigation state. When the trim adjustment is completed and the operating condition change feature value falls below the recovery threshold, the current trim state is set as the trim state corresponding to the new reference operating condition feature, and the reference operating condition feature vector is updated. This update operation enables the reference operating condition to adaptively follow changes in the ship's operating state, rather than relying on a fixed reference point, ensuring that the calculation of subsequent operating condition change feature values ​​is always based on the latest stable state.

[0073] In this embodiment, maintaining the current pitch state means not performing any new pitch adjustment action. The system enters a waiting state until the next change in operating condition characteristic exceeds the trigger threshold again, at which point a new adjustment process is triggered. This discontinuous adjustment method avoids the control burden and additional energy consumption associated with traditional continuous or high-frequency adjustments.

[0074] This embodiment achieves stable exit from trim adjustment and adaptive system calibration through dynamic updates of the hysteresis control interval and the reference operating condition characteristic vector. The hysteresis control interval suppresses frequent switching, while the update of the reference operating condition characteristic vector enables the system to follow changes in the ship's operating state. The combined effect of these two mechanisms forms a complete closed loop in the entire control process.

[0075] In a second aspect, this embodiment provides a ship trim discontinuous adjustment system based on hysteresis control, applicable to the method described in the first aspect. The system includes an operating state parameter acquisition module, an operating condition change characteristic quantity calculation module, a trend prediction module, a continuous stability duration calculation module, an energy consumption benefit assessment module, a trim adjustment execution module, and a hysteresis control monitoring module. The operating state parameter acquisition module is used to acquire the ship's current operating state parameters. The operating condition change characteristic quantity calculation module is connected to the operating state parameter acquisition module and is used to calculate the operating condition change characteristic quantity based on the operating state parameters. The trend prediction module is connected to the operating condition change characteristic quantity calculation module and is used to generate a trend prediction value based on the operating state parameters when the operating condition change characteristic quantity exceeds a preset trigger threshold. The continuous stability duration calculation module is connected to the trend prediction module and is used to determine whether the trend prediction value indicates that the operating condition change is continuous. The system calculates or resets the duration of continuous stability based on the judgment results. The energy consumption benefit assessment module is connected to the continuous stability duration calculation module. When the continuous stability duration exceeds the preset stability time threshold, it assesses the expected energy consumption benefit that can be generated by adjusting the ship's trim to different target trim states. The trim adjustment execution module is connected to the energy consumption benefit assessment module. When the expected energy consumption benefit exceeds the preset benefit threshold, it executes the trim adjustment action to adjust the ship's trim to the target trim state. The hysteresis control monitoring module is connected to the trim adjustment execution module and the operating condition change characteristic quantity calculation module. After the trim adjustment is executed, it continuously monitors the operating condition change characteristic quantity. When the operating condition change characteristic quantity falls below the preset recovery threshold, it confirms the end of this discontinuous adjustment cycle and maintains the current trim state. The recovery threshold is lower than the trigger threshold. Together, they constitute the hysteresis control range used to suppress frequent adjustments.

[0076] In this embodiment, the operational status parameter acquisition module is responsible for acquiring ship speed, draft, trim angle, main engine power, main engine speed, fuel consumption, and environmental wind, wave, and current data from shipborne sensors, the ship energy efficiency monitoring system, or the ship-shore collaborative data platform. The operational condition change characteristic quantity calculation module preprocesses the acquired raw data and calculates the Mahalanobis distance to obtain the operational condition change characteristic quantity. The trend prediction module is activated when the operational condition change characteristic quantity exceeds a trigger threshold, generating a trend prediction value through a time series prediction model. The continuous stability duration calculation module accumulates the time length from the start of the operational condition change when the trend prediction value indicates that the operational condition change is persistent. The energy consumption benefit assessment module iterates through candidate target trim states and calculates the expected energy consumption benefit when the continuous stability duration exceeds a stability time threshold using an energy consumption assessment model. The trim adjustment execution module generates an adjustment command containing a dynamic adjustment rate and sends it to the actuator when the expected energy consumption benefit exceeds a benefit threshold, while ensuring adjustment accuracy through closed-loop feedback verification. The hysteresis control monitoring module continuously monitors the characteristic quantities of the operating condition after the adjustment is executed. When the characteristic quantity falls below the recovery threshold, it confirms the end of the current discontinuous adjustment cycle and updates the baseline operating condition characteristic vector.

[0077] Each module is connected sequentially according to the data flow direction, forming a complete closed-loop control link. Among them, the hysteresis control monitoring module is connected to both the pitch adjustment execution module and the operating condition change characteristic quantity calculation module, ensuring that the operating condition change characteristic quantity can be continuously acquired and hysteresis determination can be performed after the adjustment is executed, reflecting the integrated design at the system level.

[0078] This embodiment transforms the steps described in the first aspect into an engineering-featured system architecture through modular design. Each module has clearly defined functions and interfaces, facilitating deployment and implementation in actual ship control systems. The sequentially triggered modular decision process significantly reduces the frequency of trim adjustments, avoiding unnecessary adjustments. The trend prediction module's forward-looking analysis identifies the difference between short-term disturbances and persistent changes in advance, improving the rationality of adjustment decisions. The energy consumption benefit assessment module's quantitative calculations prevent ineffective adjustments, ensuring that the additional energy consumption generated by the adjustment action itself does not offset the optimization effect. The hysteresis control monitoring module's recovery threshold setting suppresses frequent switching of the system near the threshold, improving the operational stability of the control system. The dynamic updating of the baseline operating condition feature vector enables the system to have adaptive capabilities, allowing it to follow changes in the ship's long-term operating state. This achieves a shift from continuous adjustment to conditionally triggered discontinuous adjustment, reducing the workload of actuators and improving the system's engineering feasibility.

[0079] This invention provides a method for discontinuous trim adjustment of a ship based on hysteresis control, comprising: acquiring the ship's current operating state parameters and calculating a characteristic quantity of operating condition change based on the operating state parameters; when the characteristic quantity of operating condition change exceeds a preset trigger threshold, generating a trend prediction value based on the operating state parameters; if the trend prediction value indicates that the operating condition change is persistent, accumulating the duration of sustained stability since the occurrence of the operating condition change; when the duration of sustained stability exceeds a preset stability time threshold, evaluating the expected energy consumption benefits that can be generated by adjusting the ship's trim to different target trim states; if the expected energy consumption benefits exceed a preset benefit threshold, executing a trim adjustment action to adjust the ship's trim to the target trim state; after the trim adjustment is executed, continuously monitoring the characteristic quantity of operating condition change, and when the characteristic quantity of operating condition change falls below a preset recovery threshold, confirming the end of the current discontinuous adjustment cycle and maintaining the current trim state, wherein the recovery threshold is lower than the trigger threshold, and the two together constitute a hysteresis control interval for suppressing frequent adjustments.

[0080] This invention constructs characteristic quantities of operating conditions as the basic indicators for quantifying the magnitude of operating condition changes, and sets trigger thresholds as preliminary screening conditions for entering trend prediction, thereby filtering out scenarios with insignificant operating condition changes. It uses trend prediction values ​​to determine whether operating condition changes are persistent, avoiding subsequent judgments triggered by short-term disturbances or instantaneous fluctuations. Through the cumulative calculation of continuous stable duration, it ensures that the operating condition changes are stable rather than transient processes. Through the assessment of expected energy consumption benefits, it ensures that tilt adjustment has actual energy-saving value. By setting a recovery threshold below the trigger threshold, a hysteresis control range is formed, and the current discontinuous adjustment cycle is only confirmed to have ended when the characteristic quantity of operating condition changes falls below the recovery threshold, effectively suppressing frequent triggering and recovery switching caused by operating condition fluctuations.

[0081] Compared with existing technologies, the sequential triggering criterion significantly reduces the frequency of tilt adjustments, avoiding unnecessary adjustment operations; the trend prediction mechanism improves the foresight of adjustment decisions, identifying the difference between short-term disturbances and continuous changes in advance; the energy consumption assessment mechanism avoids ineffective adjustment, preventing the additional energy consumption generated by the adjustment action itself from offsetting the optimization effect; the dynamic hysteresis control mechanism suppresses frequent switching of the system near the threshold, improving the system's operational stability; and it realizes the transformation from continuous adjustment to condition-triggered discontinuous adjustment, reducing the workload of the actuator and improving the feasibility of system engineering.

[0082] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0083] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0084] The above description is only a part of the embodiments of the present invention and does not limit the scope of protection of the present invention. Any equivalent device or equivalent process transformation made based on the content of the present invention specification and drawings, or direct or indirect application in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for discontinuous trim adjustment of a ship based on hysteresis control, characterized in that, include: Obtain the ship's current operating status parameters, and calculate the operating condition change characteristic quantity based on the operating status parameters; When the characteristic quantity of the operating condition change exceeds the preset trigger threshold, a trend prediction value is generated based on the operating status parameters; Whether the predicted trend value indicates the persistence of the change in operating conditions is determined, and the duration of continuous stability is calculated or reset based on the determination result; When the duration of sustained stability exceeds a preset stability time threshold, the expected energy savings from adjusting the ship's trim to different target trim states are assessed. If the expected energy consumption benefit exceeds the preset benefit threshold, then a trim adjustment action is performed to adjust the ship's trim to the target trim state; After the pitch adjustment is executed, the operating condition change characteristic quantity is continuously monitored. When the operating condition change characteristic quantity falls back to below the preset recovery threshold, it is confirmed that the current discontinuous adjustment cycle has ended and the current pitch state is maintained. The recovery threshold is lower than the trigger threshold, and the two together constitute a hysteresis control range for suppressing frequent adjustments.

2. The method for discontinuous trim adjustment of ships based on hysteresis control according to claim 1, characterized in that, Obtain the ship's current operating status parameters, and calculate the characteristic quantities of operating condition changes based on the operating status parameters, including: Collect data on ship speed, draft, trim angle, main engine power, main engine speed, fuel consumption, and environmental wind, wave, and current to form the original operating status parameters; Perform data preprocessing on the original operating status parameters, including: The original running state parameters are filled with missing values ​​using an interpolation algorithm to generate the filled running state parameters. Based on the Raida criterion, outlier removal processing is performed on the filled-in operating status parameters to generate outlier operating status parameters. The speed data, draft data, pitch angle data, main engine power data, main engine speed data, fuel consumption data, and environmental wind, wave and current data from different sampling frequencies in the removed operating status parameters are uniformly mapped to the same time reference, and time alignment processing is performed to generate aligned operating status parameters. The aligned running state parameters are smoothed by sliding window mean filtering or median filtering to generate smoothed running state parameters. The smoothed running state parameters are subjected to data standardization processing based on Z-score standardization or min-max standardization to generate standardized running state parameters. Standardized operating state parameters corresponding to the ship's previous stable navigation state are extracted from historical navigation data to form a baseline operating condition feature vector. Calculate the Mahalanobis distance between the standardized operating state parameters at the current moment and the baseline operating condition feature vector, and use the calculated Mahalanobis distance value as the operating condition change feature quantity.

3. The method for discontinuous trim adjustment of ships based on hysteresis control according to claim 1, characterized in that, When the characteristic quantity of the operating condition change exceeds a preset trigger threshold, a trend prediction value is generated based on the operating status parameters, including: Time series data are extracted from standardized operating status parameters within a preset time window prior to the current moment to form a historical operating status parameter sequence; A time series prediction model is constructed based on the historical operating state parameter sequence. The time series prediction model includes a differential integrated moving average autoregressive model, a long short-term memory neural network model, or a gated recurrent unit model. The time series prediction model is used to perform prediction processing on standardized operating state parameters within a future preset time window to generate a future operating state parameter sequence. The trend feature extraction process is performed on the future operating state parameter sequence relative to the standardized operating state parameters at the current time, and the extracted trend features are output as trend prediction values. The trend characteristics include the linear regression slope of the future operating state parameter sequence or the cumulative point-by-point difference between the future operating state parameter sequence and the standardized operating state parameters at the current moment.

4. The method for discontinuous trim adjustment of ships based on hysteresis control according to claim 3, characterized in that, The time series prediction model is used to perform prediction processing on standardized operating state parameters within a preset future time window to generate a future operating state parameter sequence, including: Extract continuous time window data of a preset length from the historical operating state parameter sequence. The continuous time window data contains standardized operating state parameters of the most recent N time points in the historical operating state parameter sequence. Arrange the continuous time window data in chronological order to form the model input feature vector. The model input feature vector is input into the time series prediction model, and forward inference processing is performed, including: When the time series prediction model is a differential integrated moving average autoregressive model, the autoregressive moving average calculation is performed; When the time series prediction model is a long short-term memory neural network model or a gated recurrent unit model, the forward propagation calculation of the recurrent neural network is performed step by step to obtain the model output prediction vector. The model output prediction vector contains the predicted values ​​at each time point within the future preset time window under the standardized scale. Obtain the standardization parameters recorded during the data standardization process, including the mean and standard deviation during Z-score standardization, or the minimum and maximum values ​​during min-max standardization; Based on the standardized parameters, the model output prediction vector is destandardized, and the predicted value of each standardized scale in the model output prediction vector is restored to the same physical dimension and numerical scale as the original operating state parameters, generating a destandardized prediction parameter sequence. Perform validity verification on the denormalized prediction parameter sequence, including: The predicted values ​​at each time point in the destandardized prediction parameter sequence are compared with the ship's physical operating boundary conditions, which include the upper limit of speed, the upper limit of draft, the upper limit of main engine power, and the safe range of the heel angle. If the predicted value at a certain time point in the destandardized prediction parameter sequence exceeds the corresponding ship physical operation boundary conditions, then the predicted value at that time point is corrected to the corresponding boundary value. The destandardized predicted parameter sequence after validity verification is output as the future running state parameter sequence.

5. The method for discontinuous trim adjustment of ships based on hysteresis control according to claim 3, characterized in that, Perform trend feature extraction processing on the future operating state parameter sequence relative to the standardized operating state parameters at the current time, and output the extracted trend features as trend prediction values, including: Obtain the future operating state parameter sequence, which includes standardized operating state parameters at each time point within a future preset time window; Obtain the standardized operating state parameters at the current moment, wherein the standardized operating state parameters at the current moment are the standardized operating state parameters corresponding to the current moment; A point-by-point difference calculation is performed on the standardized operating state parameters at each time point in the future operating state parameter sequence and the standardized operating state parameters at the current time to generate a parameter difference sequence. The parameter difference sequence contains the parameter difference at each time point. The parameter difference is the difference obtained by subtracting the standardized operating state parameters at the current time from the standardized operating state parameters at each time point in the future operating state parameter sequence. A linear regression fitting process is performed on the parameter difference sequence to obtain linear regression fitting parameters. The linear regression fitting parameters include slope values ​​and intercept values. The slope value represents the rate and direction of change of the future operating state parameter sequence relative to the current moment. Performing cumulative calculation processing on the parameter difference sequence includes: Calculate the algebraic sum of the parameter differences at each time point in the parameter difference sequence to obtain the cumulative difference at each point, which represents the overall change magnitude of the future operating state parameter sequence relative to the current moment; The slope value and the cumulative point-by-point difference are combined to form a trend feature vector, which includes two dimensions: the slope value and the cumulative point-by-point difference. The trend feature vector is output as the trend prediction value.

6. The method for discontinuous trim adjustment of ships based on hysteresis control according to claim 1, characterized in that, Determine whether the predicted trend value indicates a persistent change in operating conditions, and calculate or reset the duration of sustained stability based on the determination result, including: Obtain the trend prediction value, which includes a slope value and a cumulative point-by-point difference; Obtain preset persistence determination conditions, the persistence determination conditions include slope value persistence conditions and point-by-point difference accumulation persistence conditions, the slope value persistence condition is that the absolute value of the slope value is greater than a preset slope threshold, and the point-by-point difference accumulation persistence condition is that the absolute value of the point-by-point difference accumulation is greater than a preset accumulation threshold. The slope value in the trend prediction is compared with the slope value persistence condition, and the point-by-point difference accumulation in the trend prediction is compared with the point-by-point difference accumulation persistence condition. If the absolute value of the slope value is greater than the preset slope threshold, and the absolute value of the cumulative point-by-point difference is greater than the preset cumulative threshold, then the trend prediction value indicates that the change in working conditions is continuous. When it is determined that the trend prediction value indicates that the change in working conditions is continuous, the start time of the change in working conditions is obtained. The start time of the change in working conditions is the time when the characteristic quantity of the change in working conditions first exceeds the preset trigger threshold. Obtain the current time, which is the current point in time when the continuous stable duration calculation is performed; Calculate the time difference between the current time and the start time of the change in operating conditions, and output the time difference as the duration of continuous stability. The operating condition change flag is set to a continuous state, and the operating condition change flag is used to mark that the current state is a continuous operating condition change state. If the trend prediction value does not meet the persistence determination condition, then the duration of continuous stability is cleared and the operating condition change flag is set to the monitoring state.

7. The method for discontinuous trim adjustment of ships based on hysteresis control according to claim 1, characterized in that, When the duration of sustained stability exceeds a preset stability time threshold, the expected energy savings from adjusting the ship's trim to different target trim states are assessed, including: Compare the duration of sustained stability with the stability time threshold; If the duration of sustained stability is greater than the stability time threshold, then the current pitch state and the standardized operating state parameters at the current moment are obtained. Based on the standardized operating status parameters at the current moment, a set of candidate target pitch states is generated by querying the preset target pitch database to find matching candidate target pitch states. Obtain a preset energy consumption assessment model, input the current pitch state and the standardized operating state parameters at the current moment into the energy consumption assessment model, perform forward inference calculation, and obtain the predicted energy consumption of the current pitch state. Iterate through each candidate target pitch state in the candidate target pitch state set, input the candidate target pitch state and the standardized operating state parameters at the current moment into the energy consumption assessment model, perform forward inference calculation to obtain the predicted energy consumption of the candidate target pitch state, and calculate the difference between the predicted energy consumption of the current pitch state and the predicted energy consumption of the candidate target pitch state, and use the difference as the candidate expected energy consumption benefit corresponding to the candidate target pitch state; Extract the maximum value from the candidate expected energy consumption benefits corresponding to all candidate target pitch states, determine the candidate target pitch state corresponding to the maximum value as the target pitch state, and output the maximum value as the expected energy consumption benefit.

8. The method for discontinuous trim adjustment of ships based on hysteresis control according to claim 1, characterized in that, If the expected energy consumption benefit exceeds a preset benefit threshold, then a trim adjustment action is performed to adjust the ship's trim to the target trim state, including: If the expected energy consumption benefit is greater than the benefit threshold, then calculate the pitch adjustment between the current pitch state and the target pitch state; Based on the absolute value of the pitch adjustment amount, the corresponding adjustment rate parameter is queried from the preset adjustment rate mapping table to generate a pitch adjustment command containing the target pitch angle value and the adjustment rate parameter; The pitch adjustment command is sent to the pitch adjustment actuator, which then performs the pitch adjustment action. Obtain the actual pitch angle value fed back by the pitch adjustment actuator, and compare the actual pitch angle value with the target pitch state; If the deviation between the actual pitch angle value and the target pitch state is greater than or equal to a preset adjustment deviation tolerance threshold, a correction adjustment command is generated and sent to the pitch adjustment actuator again until the deviation is less than the adjustment deviation tolerance threshold.

9. The method for discontinuous trim adjustment of a ship based on hysteresis control according to claim 1, characterized in that, When the characteristic value of the operating condition change falls below the preset recovery threshold, the current discontinuous adjustment cycle is confirmed to have ended and the current pitch state is maintained, including: After the tilt adjustment action is completed, monitor the characteristic quantities of the change in the operating condition; The operating condition change characteristic quantity is compared with the recovery threshold. If the recovery threshold is lower than the preset trigger threshold, the trigger threshold and the recovery threshold together constitute the hysteresis control interval. When the characteristic value of the operating condition change falls below the recovery threshold, it is confirmed that the current discontinuous adjustment cycle has ended. Obtain the current pitch state, set the current pitch state as the pitch state corresponding to the new reference working condition feature, and update the reference working condition feature vector. Maintain the current tilt state and wait for the next change in operating condition to exceed the trigger threshold.

10. A ship trim discontinuous adjustment system based on hysteresis control, characterized in that, The system applicable to the method of any one of claims 1 to 9 comprises: The operating status parameter acquisition module is used to acquire the current operating status parameters of the ship; The operating condition change characteristic quantity calculation module is connected to the operating status parameter acquisition module and is used to calculate the operating condition change characteristic quantity based on the operating status parameters. The trend prediction module is connected to the working condition change characteristic quantity calculation module and is used to generate a trend prediction value based on the operating status parameters when the working condition change characteristic quantity exceeds a preset trigger threshold. A continuous stable duration calculation module, connected to the trend prediction module, is used to determine whether the trend prediction value indicates that the change in working conditions is continuous, and to accumulate or reset the continuous stable duration based on the determination result. An energy consumption benefit assessment module, connected to the continuous stability duration calculation module, is used to assess the expected energy consumption benefit that can be generated by adjusting the ship's trim to different target trim states when the continuous stability duration exceeds a preset stability time threshold. The trim adjustment execution module is connected to the energy consumption benefit assessment module and is used to perform trim adjustment action to adjust the ship's trim to the target trim state when the expected energy consumption benefit exceeds the preset benefit threshold. The hysteresis control monitoring module is connected to the pitch adjustment execution module and the operating condition change characteristic calculation module. It is used to continuously monitor the operating condition change characteristic after the pitch adjustment is executed. When the operating condition change characteristic falls back to below the preset recovery threshold, it confirms that the current discontinuous adjustment cycle has ended and maintains the current pitch state. The recovery threshold is lower than the trigger threshold. Together, they constitute the hysteresis control range for suppressing frequent adjustments.

Images