Ship energy efficiency optimization management and evaluation system based on big data analysis

Abstract

The application discloses a ship energy efficiency optimization management and evaluation system based on big data analysis and belongs to the technical field of ship state monitoring and energy efficiency management, which comprises an instruction synchronization module, a game processing module, a model evolution engine module, an optimization and verification loop module and a decision output module. The application adopts a double-system game architecture of parallel operation of a real ship and a high-fidelity virtual model, realizes autonomous evolution of a model to a dynamic digital mirror image by decoupling capability residual error and strategy residual error, and carries out simulation optimization and strategy verification based on the autonomous evolution. The application can accurately diagnose mechanical performance degradation and operation efficiency loss of a ship, realizes adaptive update of an evaluation benchmark and closed-loop management of an optimization decision, and significantly improves the accuracy of ship energy efficiency evaluation and the level of whole life cycle health management.

Description

Technical Field

[0001] This invention relates to the field of ship condition monitoring and energy efficiency management technology, and in particular to a ship energy efficiency optimization management and evaluation system based on big data analysis. Background Technology

[0002] Ship energy efficiency management is a core means for the modern shipping industry to achieve energy conservation, emission reduction, and lower operating costs. During actual navigation, a ship's fuel consumption is influenced by a complex interplay of factors, including the deterioration of the hull's physical condition and navigational strategies. To ensure ship energy efficiency, precise assessment and optimization of the ship's real-time energy efficiency status are necessary.

[0003] Among related technologies, Chinese invention patent CN113888088A discloses a functional verification platform for a ship energy efficiency management system based on digital twins. By receiving energy efficiency data and ship type parameters from the energy efficiency management system, it constructs a digital twin that can realistically reflect the ship's geometry, physics, navigation attitude, and navigation environment. The digital twin is used to perform virtual simulation of energy consumption during actual ship navigation. At the same time, the energy efficiency management system is used to optimize the energy efficiency of the ship's digital twin. Intelligent algorithms are integrated as a tool for effectiveness analysis and comparison. The system adopts a real-to-real and real-to-virtual dual verification method to evaluate the effect of the energy efficiency management system before and after optimization.

[0004] Regarding the aforementioned technologies, the inventors believe that although the technology can use virtual simulation to evaluate the energy efficiency optimization effect, it focuses on the functional verification at the system level, making it difficult to accurately decouple complex energy efficiency deviations and attribute them to specific physical deterioration sources or operational behaviors. Furthermore, its models are usually based on initial design parameters, making it difficult to adaptively evolve with the gradual deterioration of ship performance over time, thus limiting the accuracy of the evaluation of ships in service. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a ship energy efficiency optimization management and evaluation system based on big data analysis. It employs a high-fidelity virtual model that operates in parallel with the actual ship and performs two-way expected game processing to decompose energy efficiency deviations. Simultaneously, it drives the model to autonomously evolve into a dynamic digital mirror image. This system can accurately test and distinguish the impact of ship physical performance degradation and operational strategy advantages and disadvantages on energy efficiency, thereby enabling closed-loop optimization verification and predictive maintenance decisions.

[0006] To achieve the above objectives, this application adopts the following technical solution: Firstly, a ship energy efficiency optimization management and evaluation system based on big data analysis is provided, including: a command synchronization module, used to acquire navigation control commands from the actual ship and synchronously send the navigation control commands to a preset high-fidelity virtual model used to simulate the ideal physical characteristics of the ship, driving the high-fidelity virtual model to run in parallel with the actual ship and generating parallel operation data; and a game processing module, used to perform two-way expected game processing based on the parallel operation data, to obtain the capability residual by calculating the first difference between the actual energy consumption of the actual ship and the theoretical expected energy consumption output by the high-fidelity virtual model, and to calculate the second difference between the actual power of the actual ship's main engine and the theoretical optimal power derived from the high-fidelity virtual model. The system is designed to obtain policy residuals; the model evolution engine module is used to iteratively adjust the core physical parameters in the high-fidelity virtual model based on the capability residuals, so that the high-fidelity virtual model evolves into a dynamic digital image representing the current physical state of the actual ship; the optimization and verification loop module is used to perform simulation optimization using the evolved dynamic digital image, generate an optimized policy package with the goal of reducing policy residuals, and perform two-way expected game processing to perform closed-loop verification of the optimized policy package to obtain the verified optimized policy; the decision output module is used to generate a comprehensive evaluation report and trigger corresponding maintenance or execution instructions based on the parameters, capability residuals, policy residuals and verified optimized policies of the dynamic digital image.

[0007] Based on the above technical solution, the ship energy efficiency optimization management and evaluation system based on big data analysis provided in this application adopts a high-fidelity virtual model that runs in parallel with the actual ship and performs two-way expected game processing to decompose energy efficiency deviation. At the same time, it drives the model to autonomously evolve into a dynamic digital mirror. This means that the system can accurately test and distinguish the impact of ship physical performance degradation and the advantages and disadvantages of operating strategies on energy efficiency, and then carry out closed-loop optimization verification and predictive maintenance decisions.

[0008] In conjunction with the first aspect mentioned above, in one possible implementation, the game processing module includes: a forward comparison unit, used to extract the theoretical expected energy consumption and actual energy consumption from the parallel operation data, perform forward comparison processing, calculate and output the capability residual; a reverse comparison unit, used to extract the actual water speed reached by the actual ship from the parallel operation data, and drive the high-fidelity virtual model to back-calculate the theoretical optimal power based on the water speed, while obtaining the actual power of the actual ship's main engine, performing reverse comparison processing, calculating and outputting the strategy residual; and a structured storage unit, used to establish the correlation between the capability residual and the strategy residual and the current navigation condition label, forming structured game difference data.

[0009] In conjunction with the first aspect mentioned above, in one possible implementation, the model evolution engine module includes: a pattern recognition unit, used to perform pattern recognition processing on the capability residuals, distinguishing between a first type of residual pattern caused by changes in hull resistance and a second type of residual pattern caused by changes in propulsion efficiency; a parameter mapping unit, used to map the first type of residual pattern to an adjustment amount for the hull resistance parameter in the high-fidelity virtual model, and to map the second type of residual pattern to an adjustment amount for the propeller efficiency parameter in the high-fidelity virtual model; and an iterative correction unit, used to iteratively correct the hull resistance parameter and the propeller efficiency parameter based on the adjustment amounts of the hull resistance parameter and the propeller efficiency parameter, thereby completing the autonomous evolution of the high-fidelity virtual model into a dynamic digital image.

[0010] In conjunction with the first aspect mentioned above, in one possible implementation, the pattern recognition unit includes: a linear recognition subunit, used to calculate the first derivative of the capability residual over time to identify a linear growth trend, and to calculate the correlation coefficient between the capability residual and environmental parameters to eliminate interference from environmental factors, thereby identifying a first type of residual pattern; and a correlation analysis subunit, used to subtract the first type of residual pattern from the capability residual to obtain the residual fluctuation component, and to calculate the correlation strength between the residual fluctuation component and the host operating parameters, thereby identifying a second type of residual pattern.

[0011] In conjunction with the first aspect mentioned above, in one possible implementation, the optimization and verification loop module includes: a simulation optimization unit, used to perform combined optimization of airspeed setting, heading, and wing deflection angle in a dynamic digital mirror, searching for parameter combinations that can optimize the predicted strategy residuals to form an initial optimization strategy package; a strategy verification unit, used to use the initial optimization strategy package as a new navigation control command to drive the execution of two-way expected game processing to obtain the strategy residuals for verification; and a threshold determination unit, used to determine whether the strategy residuals for verification meet a preset optimization effectiveness threshold. If yes, the initial optimization strategy package is marked as a verified optimization strategy; otherwise, the initial optimization strategy package is discarded and a new round of simulation optimization is triggered.

[0012] In conjunction with the first aspect mentioned above, in one possible implementation, the simulation optimization unit includes: a constraint acquisition subunit, used to acquire the expected arrival time constraint and the carbon emission intensity index constraint required by regulations in the voyage plan; a boundary processing subunit, used to treat the expected arrival time constraint as a hard boundary condition for combinatorial optimization and the carbon emission intensity index constraint as an optimization objective in multi-objective optimization calculation; and a solution execution subunit, used to solve the multi-objective optimization calculation under the premise of satisfying the hard boundary conditions and generate an initial optimization strategy package.

[0013] In conjunction with the first aspect mentioned above, in one possible implementation, the system also includes a model building module: an ideal parameter acquisition unit, used to acquire ideal physical parameters characterizing the performance of the ship during the design phase, wherein the ideal physical parameters include the smooth hull resistance curve, the open water characteristic curve of the new propeller, and the standard efficiency spectrum of the main engine; and a simulation building unit, used to construct a high-fidelity virtual model capable of simulating and calculating the theoretical expected energy consumption and theoretical optimal power based on the ideal physical parameters and the ship's kinematics equations.

[0014] In conjunction with the first aspect mentioned above, in one possible implementation, the decision output module includes: a state quantification unit, used to quantify the evolved core physical parameters in the dynamic digital image into a ship physical condition deterioration certificate; a diagnosis generation unit, used to analyze the contribution ratio of capability residuals and strategy residuals in the total energy consumption deviation and generate an energy efficiency performance diagnosis report; and a maintenance triggering unit, used to automatically generate a predictive maintenance work order containing the ship physical condition deterioration certificate and the energy efficiency performance diagnosis report when the long-term trend value of the capability residuals exceeds a preset maintenance threshold.

[0015] In conjunction with the first aspect mentioned above, in one possible implementation, the system also includes a digital history module: a sequence recording unit, used to continuously record the values ​​and corresponding timestamps of the core physical parameters of the model evolution engine after each iteration adjustment, forming a parameter evolution sequence; an association construction unit, used to associate the parameter evolution sequence with the navigation logs and maintenance records of the same period to construct a ship performance degradation timeline; and a trend prediction unit, used to generate a prediction result of the future energy efficiency status based on the performance degradation timeline and using a trend prediction model.

[0016] Secondly, it provides a method for ship energy efficiency optimization management and evaluation based on big data analysis, including: acquiring navigation control commands from the actual ship and synchronously sending the navigation control commands to a pre-set high-fidelity virtual model used to simulate the ideal physical characteristics of the ship; driving the high-fidelity virtual model and the actual ship to run in parallel and generating parallel operation data; performing two-way expected game processing based on the parallel operation data; obtaining the capacity residual by calculating the first difference between the actual energy consumption of the actual ship and the theoretical expected energy consumption output by the high-fidelity virtual model; and calculating the actual power of the actual ship's main engine and the theoretical optimal power derived from the high-fidelity virtual model. The second difference between the rates yields the strategy residual. Based on the capability residual, the core physical parameters in the high-fidelity virtual model are iteratively adjusted to evolve the high-fidelity virtual model into a dynamic digital image representing the current physical state of the actual ship. The evolved dynamic digital image is used for simulation optimization to generate an optimized strategy package with the goal of reducing the strategy residual. A two-way expected game is then performed to perform closed-loop verification of the optimized strategy package, resulting in a verified optimized strategy. Based on the parameters, capability residual, strategy residual, and verified optimized strategy of the dynamic digital image, a comprehensive evaluation report is generated and corresponding maintenance or execution instructions are triggered.

[0017] Compared with the prior art, the present invention has the following advantages: This invention, through bidirectional decoupling of energy efficiency deviations, can accurately distinguish and quantify the capability residuals caused by physical performance degradation such as hull fouling and propeller aging, as well as the strategy residuals caused by improper operational behaviors such as navigation strategies and main engine control. This precise attribution diagnosis allows managers to take targeted measures, such as arranging hull cleaning or optimizing navigation plans, thereby improving the accuracy of energy efficiency problem diagnosis and the effectiveness of solutions.

[0018] This invention constructs an adaptive dynamic digital mirror, using a model evolution engine to continuously iteratively correct the core physical parameters of a high-fidelity virtual model based on capability residuals. This enables the benchmark model used for evaluation and testing to reflect the ship's current real physical state in real time, overcoming the shortcomings of traditional methods that use static sea trial data or initial models as benchmarks, which lead to inaccurate evaluation results over time. This ensures the continuous effectiveness and high reliability of energy efficiency assessment and testing throughout the ship's entire lifecycle.

[0019] This invention establishes a feedforward control loop from simulation optimization to closed-loop verification. It actively searches for and generates optimization strategy packages using a highly simulated dynamic digital mirror, and pre-verifies the energy-saving effects of the strategies in a virtual environment. This "calculate first, then act" approach avoids the risks of directly applying unverified strategies to real ships, ensuring that the recommended optimization measures effectively reduce strategy residuals. This elevates ship energy efficiency management from passive analysis and diagnosis to proactive, predictable optimization control.

[0020] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0021] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating the ship energy efficiency optimization management and evaluation method based on big data analysis provided in the embodiments of this application; Figure 2 The structural architecture diagram of the ship energy efficiency optimization management and evaluation system based on big data analysis provided in the embodiments of this application; Figure 3 This is a capability residual evolution identification curve provided in the embodiments of this application.

[0023] Figure 4 This is a distribution diagram verifying the optimization strategy provided in the embodiments of this application. Detailed Implementation

[0024] In the description of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. The "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" means one or more, and "multiple" means two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0025] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0026] The ship energy efficiency optimization management and evaluation system based on big data analysis provided in this application embodiment can be applied to, for example... Figure 1 The ship energy efficiency optimization management and evaluation method based on big data analysis shown in the example is as follows: Figure 1 As shown, the method includes: Obtain navigation control commands from the actual ship and synchronously send the navigation control commands to a preset high-fidelity virtual model used to simulate the ideal physical characteristics of the ship, drive the high-fidelity virtual model to run in parallel with the actual ship, and generate parallel operation data; Based on parallel operation data, a two-way expected game is performed. The capability residual is obtained by calculating the first difference between the actual energy consumption of the actual ship and the theoretical expected energy consumption output by the high-fidelity virtual model. The strategy residual is obtained by calculating the second difference between the actual power of the main engine of the actual ship and the theoretical optimal power derived by the high-fidelity virtual model. Based on the capability residuals, the core physical parameters in the high-fidelity virtual model are iteratively adjusted, so that the high-fidelity virtual model evolves into a dynamic digital image representing the current physical state of the real ship. The evolved dynamic digital mirror is used for simulation optimization. An optimized policy package is generated with the goal of reducing the policy residual. The optimized policy package is then verified by two-way expected game processing to obtain the verified optimized policy. Based on the parameters, capability residuals, policy residuals, and verified optimization strategies of the dynamic digital image, a comprehensive evaluation report is generated and corresponding maintenance or execution instructions are triggered.

[0027] like Figure 2 As shown in the embodiments of this application, a ship energy efficiency optimization management and evaluation system based on big data analysis is provided, including: The command synchronization module is used to acquire the navigation control commands of the actual ship and synchronously send the navigation control commands to the preset high-fidelity virtual model used to simulate the ideal physical characteristics of the ship, drive the high-fidelity virtual model to run in parallel with the actual ship, and generate parallel operation data. The game processing module is used to perform two-way expected game processing based on parallel operation data. It obtains the capability residual by calculating the first difference between the actual energy consumption of the actual ship and the theoretical expected energy consumption output by the high-fidelity virtual model, and obtains the strategy residual by calculating the second difference between the actual power of the actual ship's main engine and the theoretical optimal power derived from the high-fidelity virtual model. The model evolution engine module is used to iteratively adjust the core physical parameters in the high-fidelity virtual model based on the capability residual, so that the high-fidelity virtual model evolves into a dynamic digital image that represents the current physical state of the real ship. The optimization and verification loop module is used to perform simulation optimization using the evolved dynamic digital mirror, generate an optimized policy package with the goal of reducing policy residuals, and perform closed-loop verification of the optimized policy package by performing two-way expected game processing to obtain the verified optimized policy. The decision output module is used to generate a comprehensive evaluation report and trigger corresponding maintenance or execution instructions based on parameters, capability residuals, strategy residuals, and verified optimization strategies from a dynamic digital image.

[0028] It should be noted that a closed-loop control framework based on parallel system theory is constructed. First, a command synchronization module drives a high-fidelity virtual model representing the ship's ideal performance to operate synchronously with the actual ship, establishing a parallel reference system between reality and virtuality. An innovative two-way expected game analysis is executed using a game processing module. A forward comparison of actual energy consumption and theoretical expected energy consumption yields the capability residual, quantifying the deterioration of the ship's physical state; a reverse comparison of actual power and theoretical optimal power yields the strategy residual, quantifying the merits of operational strategies. This achieves precise decoupling and attribution of total energy efficiency deviations. Based on the capability residuals, a model evolution engine iteratively adjusts the core physical parameters of the virtual model, evolving it into a dynamic digital mirror that truly reflects the current ship condition. This evolved dynamic digital mirror is used for simulation optimization, generating optimized strategies with the goal of reducing strategy residuals. The effectiveness of these strategies is verified through an internal closed loop, ultimately outputting comprehensive evaluation and decision commands.

[0029] In one possible implementation of the embodiments of this application, combined with Figure 2 The game processing module includes: The forward comparison unit is used to extract the theoretical expected energy consumption and actual energy consumption from the parallel operation data, perform forward comparison processing, calculate and output the capability residual; The reverse comparison unit is used to extract the actual water speed reached by the real ship from the parallel operation data, and drive the high-fidelity virtual model to back-calculate the theoretical optimal power based on the water speed. At the same time, it obtains the actual power of the real ship's main engine, performs reverse comparison processing, calculates and outputs the strategy residual. Structured storage units are used to establish the correlation between capability residuals and strategy residuals and the current navigation condition label, forming structured game difference data.

[0030] In some implementations, the game theory processing module serves as the core computational component, precisely quantifying the deviations of the actual ship from its ideal state in terms of physical performance and operational strategies. Through a bidirectional comparison and inversion calculation process, the comprehensive energy efficiency deviation is decomposed into capability residuals attributed to the ship's equipment condition and strategy residuals attributed to navigation and maneuvering behavior. The module includes a forward comparison unit to quantify the additional energy consumption caused by physical degradation such as hull fouling and propeller wear. This unit's operation is triggered within a preset calculation cycle, typically 5 to 15 minutes. The actual energy consumption of the actual ship within the current calculation cycle is extracted from the parallel operation data generated by the command synchronization module. This data comes from the integral values ​​of the ship's onboard fuel flow meter or power management system. Simultaneously, the theoretical expected energy consumption output by the high-fidelity virtual model under the same navigation control commands and external environment is extracted. The capability residual is calculated using the following formula. : ; in, Represents the residual capacity, measured in kilowatt-hours per hour or kilograms per hour, and characterizes the average power loss due to the deterioration of physical properties; and They are in the time window The actual total energy consumption and the theoretical total energy consumption within the system; It is the calculation period. A continuously positive and increasing capability residual. This directly indicates the ship's physical performance, such as the continuous deterioration of hull resistance. The module includes a backcomparison unit to decouple and quantify power waste caused by improper pilot operation or autopilot settings. This unit obtains the actual average speed of the ship across the water within the current calculation period from parallel operation data. This data is precisely measured by an onboard Doppler log or electromagnetic log to eliminate the influence of ocean currents. This unit will then measure the actual surface speed. As input, the high-fidelity virtual model is driven to perform inverse solving, calculating the theoretically optimal main engine power required to overcome the current environmental forces at that speed. Simultaneously, the unit obtains the actual output power of the host computer within the same time period from the host monitoring system. The strategy residuals are calculated using the following formula B. : ; in, The residual represents the strategy, expressed in kilowatts, and directly reflects the power consumption exceeding the theoretical optimum. This is the actual measured power of the main unit; This is the theoretically optimal power calculated by the model inverse calculation. A significantly positive policy residual. In engineering terms, this indicates room for optimization in current navigation strategies, such as speed fluctuation control, load and trim adjustments, and rudder effectiveness utilization. The module also includes a structured storage unit for building a game-theoretic difference database for big data analysis and trend prediction. This unit will forward compare the capability residuals output by the unit. Strategy residuals of the inverse comparison unit output The system forcibly associates the current navigation status with a set of navigation condition labels describing the current navigation state. This label set is high-dimensional and includes key parameters such as timestamp, ship latitude and longitude, surface speed, main engine speed, heading, draft, wind speed and direction, and wave height and period. By storing the residual obtained from each calculation and its corresponding condition label as a data record, structured game difference data is formed, providing a traceable and analyzable data foundation for subsequent model evolution, strategy optimization, and energy efficiency assessment.

[0031] For example, the game theory processing module processes the energy efficiency data of a 300,000-ton oil tanker on a specific voyage, with a preset calculation period. The time interval is 0.25 hours, or 15 minutes. The forward comparison unit obtains the actual total energy consumption of the actual ship during this time period through the command synchronization module. The weight is 875.0 kg. Simultaneously, the theoretical total energy consumption of the high-fidelity virtual model under the same commands and environment is extracted. It is 812.5 kg. Substitute it into the formula. Calculate the capability residual Kilograms per hour, this consistently positive value quantifies the power loss due to hull fouling; subsequently, the reverse comparison unit obtains the ship's average water speed during that period. The speed is set at 14.5 knots, and the virtual model is used to solve in reverse for the theoretically optimal main engine power required to reach that speed. The power output is 22,500 kilowatts, combined with the actual output power obtained from the main unit monitoring system during the same period. It is 23,200 kilowatts. Substitute it into the formula. Calculate the strategy residual The kilowatts clearly indicate the amount of power wasted due to suboptimal navigation strategies. Finally, the structured storage unit forcibly associates and stores the calculated capacity residual of 250.0 kg / h and strategy residual of 700 kilowatts with high-dimensional navigation condition labels, including the current timestamp, draft of 12.5 meters, and wind speed of 10.2 meters per second, forming traceable structured game difference data.

[0032] In one possible implementation, combining Figure 2 The model evolution engine module includes: The pattern recognition unit is used to perform pattern recognition processing on the capability residuals, distinguishing between the first type of residual patterns caused by changes in hull resistance and the second type of residual patterns caused by changes in propulsion efficiency. The parameter mapping unit is used to map the first type of residual mode to the adjustment amount of the hull resistance parameter in the high-fidelity virtual model, and to map the second type of residual mode to the adjustment amount of the propeller efficiency parameter in the high-fidelity virtual model. The iterative correction unit is used to iteratively correct the hull resistance parameter and the propeller efficiency parameter based on the adjustment amount of the hull resistance parameter and the propeller efficiency parameter, so as to complete the autonomous evolution of the high-fidelity virtual model into a dynamic digital image.

[0033] In some implementations, the model evolution engine module is the core of achieving adaptive and self-learning functions. It enables the autonomous evolution of a high-fidelity virtual model by continuously iteratively adjusting core physical parameters, transforming it from a static benchmark representing ideal performance during the design phase into a dynamic digital mirror capable of accurately reproducing the current physical state of the actual ship. The pattern recognition unit within the module decouples two different physical causes of degradation from the comprehensive capability residuals output by the game processing module. This unit processes the time-series data of the capability residuals over a relatively long timescale, such as 24 to 72 hours. By applying low-pass filtering or moving average algorithms, it extracts the slowly monotonically increasing trend component in the capability residuals. This component is identified as a first-type residual pattern caused by the continuous increase in hull resistance, such as marine organism attachment and changes in hull surface roughness. Subtracting the identified first-type residual pattern from the original capability residual sequence yields a more volatile residual component more closely related to changes in main engine operating conditions. This is identified as a second-type residual pattern caused by a decrease in propulsion system efficiency, such as propeller damage or suboptimal operation. The parameter mapping unit establishes a quantitative mapping relationship between identified residual patterns and specific physical parameter adjustments in the high-fidelity virtual model. This unit receives first-type and second-type residual patterns as input. The first-type residual pattern is mapped to an adjustment for the hull resistance parameter in the high-fidelity virtual model using a preset empirical function or lookup table. This is typically represented by an increment in the dimensionless hull resistance fouling coefficient. Similarly, this unit maps the second-type residual pattern to an adjustment for the propeller efficiency parameter in the high-fidelity virtual model, typically represented by an increment in the efficiency reduction coefficient used to correct the propeller open-water characteristic curve. The iterative correction unit performs specific model parameter update operations, completing the convergence process from the high-fidelity virtual model to the dynamic digital image. This unit is triggered at a fixed update cycle, such as once daily, or when the accumulated capability residual reaches a preset threshold. Based on the adjustment output from the parameter mapping unit, the core physical parameters of the model are iteratively corrected. This process can be characterized by the following formula: ; ; in, and These are the hull resistance parameters before and after the update; and These are the propeller efficiency parameters before and after the update. and It is a dimensionless adjustment amount calculated by the parameter mapping unit. and This is the learning rate or iterative gain coefficient, typically ranging from 0.01 to 0.1. It controls the correction step size, ensuring the stability and convergence of the model evolution. Through this series of controlled iterative corrections, the high-fidelity virtual model gradually approximates the actual physical performance of the ship, completing its autonomous evolution into a dynamic digital mirror. Figure 3 As shown in the figure, the thin solid line represents the capability residual of the original acquisition. The first type of residual pattern extracted by the pattern recognition unit shows a clear linear increase, and this trend quantifies the impact of the ship's resistance deteriorating over time on energy efficiency.

[0034] For example, the model evolution engine module updates the model parameters of a bulk carrier. The pattern recognition unit extracts the long-term trend component of the capacity residual as 120.0 kW over a 72-hour timescale, identifying it as an increase in hull resistance caused by hull fouling, i.e., the first type of residual mode. It then calculates the remaining fluctuation component as 35.0 kW, identifying it as the second type of residual mode caused by a decrease in propeller performance. Subsequently, the parameter mapping unit maps the first type of residual mode to the hull resistance parameter adjustment amount using an empirical function. The value is 0.05, which maps the second type of residual mode to the propeller efficiency parameter adjustment. The learning rate is -0.02; the final iteration correction unit sets the learning rate when triggering the update cycle. and All values ​​are 0.1; retrieve the resistance parameters before the update. For 1.05 and efficiency parameters The value is 0.98. Substituting this into the formula: And the formula: Through the above calculations, the core physical parameters are iteratively corrected, enabling the high-fidelity virtual model to gradually approach the actual performance of the ship and evolve into a dynamic digital image.

[0035] In one possible implementation, combining Figure 2 The pattern recognition unit includes: The linear identification subunit is used to calculate the first derivative of the capability residual over time to identify a linear growth trend, and to calculate the correlation coefficient between the capability residual and environmental parameters to eliminate interference from environmental factors, thereby identifying the first type of residual pattern. The correlation analysis subunit is used to subtract the first type of residual pattern from the capability residual to obtain the residual fluctuation component, and to calculate the correlation strength between the residual fluctuation component and the host operating parameters, thereby identifying the second type of residual pattern.

[0036] In some implementations, the implementation details of the pattern recognition unit are crucial to the accuracy of the model evolution engine. Through sophisticated data analysis algorithms, the composite capability residuals are decomposed and attributed to two independent physical degradation sources: the hull and the propulsion system. A linear recognition sub-unit within the module is used to accurately identify and quantify the energy consumption increment caused by the cumulative effect of hull fouling. This sub-unit processes a long-period, typically 15 to 30-day, time-series data of capability residuals. The first derivative of the capability residual sequence is calculated using the following formula to capture its growth trend: ; in, The growth rate of the representative capability residual. For the capacity residual time series, For time. When The value remains positive and fluctuates little over a relatively long period; for example, its average value within an observation window exceeds a preset minimum positive threshold, initially indicating a linear growth trend. To eliminate spurious trends caused by environmental factors, such as prolonged navigation in harsh sea conditions, this sub-unit further calculates the capability residual. The correlation coefficient with concurrent environmental parameters, especially wind speed and wave height, such as the Pearson correlation coefficient, is used. If this correlation coefficient is lower than a set threshold, such as 0.3, the linear growth trend is determined to originate from the hull itself, thus identifying and outputting it as a Type I residual pattern. The correlation analysis subunit extracts the portion related to propulsion system performance changes from the remaining residual data. This subunit performs a subtraction operation, using the following formula to separate the identified Type I residual pattern from the original capability residuals, obtaining the residual fluctuation component: ; in, It is the residual fluctuation component. It is the original ability residual. This is the first type of residual pattern identified by the linear identification subunit. This residual fluctuation component... Physically, this represents the energy consumption deviation caused by factors other than long-term hull deterioration. This sub-unit calculates this residual fluctuation component. The correlation strength between the component and the time series of main engine operating parameters, such as main engine speed, torque, or slip rate, within the same time period is typically determined using algorithms such as cross-correlation analysis or dynamic time warping. When the calculated correlation strength exceeds a high preset threshold, such as 0.7, the fluctuation component is determined to be highly correlated with the operating state or efficiency changes of the propulsion system, thus identifying it and outputting it as a Type II residual mode. This step effectively distinguishes between propulsion efficiency declines caused by increased propeller surface roughness, blade damage, or operation under off-design conditions.

[0037] For example, the pattern recognition unit identifies the energy efficiency degradation of a very large crude carrier over a 30-day voyage, while the linear recognition subunit obtains the time series of the capability residuals within that period. Substitute into the formula: Calculate the growth rate The average value is 0.15 kilowatt-hours, exhibiting a consistently positive characteristic, followed by calculation. The correlation coefficient with the average wave height and wind speed during the same period was 0.12. Since this was below the set threshold of 0.3, the linear increase was determined to be caused by hull fouling, and the first type of residual pattern was identified and output. The power is 45.0 kilowatts; then, the formula is executed for the correlation analysis sub-unit: The residual fluctuation component is obtained by separating the first type of mode from the current capability residual. The value of the component was calculated to be 10.0 kW, and the cross-correlation strength between the component and the main engine slip rate parameter sequence was calculated to be 0.85. Since it exceeded the preset threshold of 0.7, it was determined to be highly correlated with the decline in propulsion system efficiency. Thus, the second type of residual mode was identified and output as 10.0 kW, achieving accurate attribution and decoupling of long-term hull deterioration and short-term propeller efficiency decline.

[0038] In one possible implementation, combining Figure 2 The optimization and verification loop module includes: The simulation optimization unit is used to perform combined optimization of air speed setting, heading and wing deflection angle in dynamic digital mirror, and search for parameter combinations that can optimize the predicted strategy residuals to form an initial optimization strategy package. The strategy verification unit is used to take the initial optimized strategy package as a new navigation control command, drive the execution of two-way expected game processing, and obtain the strategy residual for verification. The threshold determination unit is used to determine whether the strategy residual used for verification meets the preset optimization effectiveness threshold. If yes, the initial optimization strategy package is marked as a verified optimization strategy; otherwise, the initial optimization strategy package is discarded and a new round of simulation optimization is triggered.

[0039] In some implementations, the optimization and verification loop module constitutes an automated decision support system for feedforward control and closed-loop verification. Utilizing an evolved dynamic digital mirror, the system proactively simulates and optimizes the navigation strategy, virtually verifying the effectiveness of the optimization results and generating executable instructions that significantly reduce strategy residuals. The simulation optimization unit within the module explores and determines a set of energy-optimal operating parameter combinations without violating the core constraints of the navigation mission. This unit uses the evolved dynamic digital mirror as the simulation platform, which accurately represents the ship's current physical state. Speed ​​settings, heading adjustments, and the deflection angles of potential auxiliary propulsion devices such as wind deflectors are used as optimization variables to construct a multi-dimensional solution space. Optimization algorithms, such as genetic algorithms or particle swarm optimization algorithms, are used for efficient searching within this space. The objective function of the algorithm is to minimize the predicted strategy residual; that is, in the simulation, for each parameter combination, the difference between its power consumption and the theoretically optimal power required to reach the corresponding speed is calculated. After a certain number of iterative calculations, typically hundreds to thousands of simulations, the optimization unit outputs a set of parameter combinations that minimize the residual of the predicted strategy, forming an initial optimized strategy package covering the next few hours of voyage. The strategy verification unit quantifies the actual energy-saving effect of the initial optimized strategy package in a controlled virtual environment to avoid the risk of directly applying an unverified strategy to a real ship. This unit uses the initial optimized strategy package generated in the previous step as a new set of virtual navigation control commands and drives the entire system to perform a complete two-way expected game. These commands are then input into a dynamic digital mirror to simulate the ship's response, obtaining the simulated actual power. And the simulated actual water speed At the same time, The theoretical optimal power at that speed is obtained by inputting the data into a high-fidelity virtual model. The difference between these two values ​​is calculated to obtain the policy residual specifically for verification. The threshold determination unit makes the final decision on the effectiveness of the optimized policy. This unit receives the policy residual for verification calculated by the policy verification unit and compares it with a preset optimization effectiveness threshold. This process can be represented by the following formula: ; in, This represents the policy residual used for verification, and this value is output by the policy verification unit. This is a preset optimization effectiveness threshold, a key engineering parameter. Its value is typically set as a percentage of the current average policy residual; for example, it may require the optimized policy to achieve at least a 2% power saving. If... Less than If the optimization is deemed effective, the initial optimization strategy package is immediately marked as a validated optimization strategy and prepared for output to the decision-making system. If... Not less than If this happens, it means the energy-saving effect of the optimization scheme has not met the expected standard, the initial optimization strategy package will be discarded, and the system will trigger the simulation optimization unit, which may adjust the initial conditions or constraints of the optimization algorithm and start a new round of simulation optimization, forming a closed loop of continuous improvement. Figure 4 As shown, solid dots represent qualified parameter combinations that fall below the optimization effectiveness threshold. This closed-loop verification mechanism ensures that the final verified optimization strategy can achieve the expected power savings in the actual physical model.

[0040] For example, the optimization and verification loop module optimizes the navigation strategy of a 10,000-ton cargo ship. The simulation optimization unit uses the evolved dynamic digital image as the simulation platform and performs thousands of iterative simulations in a multi-dimensional solution space consisting of speed setting, heading, and wing deflection angle using a genetic algorithm. It outputs a set of parameter combinations that minimizes the residual of the predicted strategy to form an initial optimized strategy package. Subsequently, the strategy verification unit inputs this initial optimized strategy package as a virtual control command into the dynamic digital image to simulate the actual power. For 18,500 kilowatts and water speed The speed was set at 14.2 knots, and this speed was input into a high-fidelity virtual model to deduce the theoretically optimal power. The power is 18,200 kilowatts. The difference between the two is used to obtain the strategy residual for verification. The current average strategy residual is 300 kW; the final threshold determination unit obtains the current average strategy residual as 500 kW, and sets the optimization effectiveness threshold according to the principle of saving at least 10% of power. The value is 450 kilowatts. Substituting this into the formula: The determination is based on 300 kilowatts. The result was 450 kilowatts. The conclusion was that the preset optimization effectiveness threshold was met. Therefore, the initial optimization strategy package was marked as a verified optimization strategy. If the calculation result did not meet the inequality, the strategy package was discarded and a new round of simulation optimization was triggered.

[0041] In one possible implementation, combining Figure 2 The simulation optimization unit includes: The constraint acquisition subunit is used to acquire the expected arrival time constraints and the carbon emission intensity indicators required by regulations in the voyage plan; The boundary processing subunit is used to treat the expected arrival time constraint as a hard boundary condition for combinatorial optimization and the carbon emission intensity index constraint as the optimization objective in multi-objective optimization calculation. Solve the execution sub-unit, which is used to solve multi-objective optimization calculations under the premise of satisfying hard boundary conditions and generate an initial optimization strategy package.

[0042] In some implementations, the internal implementation of the simulation optimization unit is crucial to ensuring the feasibility and compliance of the optimization strategy in the real world. To ensure that the simulation optimization process not only pursues theoretical energy efficiency optimization but also strictly adheres to the environmental requirements of the voyage's commercial constraints, the unit actively interfaces with the ship's voyage management system and built-in regulatory database through its internal constraint acquisition subunit. This allows for the real-time acquisition of two core external constraints: the estimated arrival time constraint, a clearly defined timestamp, is parsed from the voyage plan; and the current mandatory carbon emission intensity index constraint, typically measured and rated in grams of carbon dioxide per ton-nautical mile per year, is retrieved from the regulatory database. The boundary processing subunit precisely converts these business-level constraints into mathematical boundaries recognizable by the optimization algorithm. This unit defines the acquired estimated arrival time constraint as a hard boundary condition in the combinatorial optimization process. This means that any parameter combination that causes the predicted total voyage time to exceed this timestamp will be directly deemed infeasible and discarded during the optimization process. Meanwhile, this unit uses carbon emission intensity as a parallel optimization objective, which, together with the objective of reducing strategy residuals, constitutes a multi-objective optimization problem. This allows the system to balance energy consumption per unit time with carbon emissions per unit of transport work while ensuring on-time arrival. The solution execution subunit solves this multi-objective optimization problem within the defined constraint framework, generating an initial optimization strategy package that balances economy and compliance. This subunit employs a multi-objective optimization algorithm, such as the non-dominated sorting genetic algorithm NSGA-II with an elitist strategy, to find a set of Pareto optimal solutions. Its solution process can be abstracted as the following optimization problem: ; ; in, It is the combination of parameters to be optimized, including variables such as speed and heading. It is a multi-objective function, where It is the policy residual function predicted based on dynamic digital mirroring. It is a function representing the predicted carbon emission intensity. It is based on parameter combinations The calculated total sailing time function, and This is the estimated arrival time, which serves as a hard boundary condition. The solution involves determining the execution sub-unit while satisfying the time constraints. Less than or equal to Among all solutions, find the one that makes and The solution set is made as small as possible, and the solution with the best overall performance is selected from it and packaged into an initial optimization strategy package.

[0043] For example, the simulation optimization unit optimizes the strategy of a large container ship during a transoceanic voyage, and the constraint acquisition subunit parses the estimated arrival time constraints from the voyage management system. To ensure the flight time does not exceed 120 hours, and to obtain carbon emission intensity index constraints and rating requirements from the regulatory database, the boundary processing subunit sets 120 hours as a hard boundary condition and treats carbon emission intensity as a multi-objective parallel to energy efficiency optimization. Finally, the solution execution subunit uses the NSGA-II algorithm to solve the optimization problem. Among them, the combination of parameters to be optimized Including variables such as speed and heading, if a certain combination of parameters produces a predicted total sailing time A time limit of 125 hours or more than 120 hours is considered an infeasible solution and discarded. However, solutions that meet the requirements... In the solution set, by balancing the prediction strategy residual function For example, a calculated value of 320 kilowatts and a predicted carbon emission intensity index function The best-performing unpacking strategy package is selected to form the initial optimization strategy package.

[0044] In one possible implementation, combining Figure 2 The system also includes a model building module: The ideal parameter acquisition unit is used to acquire ideal physical parameters characterizing the performance of the ship during the design phase. These ideal physical parameters include the smooth hull resistance curve, the new propeller open water characteristic curve, and the main engine standard efficiency spectrum. The simulation building unit is used to construct a high-fidelity virtual model based on ideal physical parameters and ship motion equations, capable of simulating and calculating theoretically expected energy consumption and theoretically optimal power.

[0045] In some implementations, the model building module serves as the initial foundation and reference system for the entire system. Based on the theoretical performance data of the ship during the design and construction phases, it constructs a high-fidelity virtual model representing the ship's optimal, non-deteriorating state in a single operation. This is typically executed during the initial deployment on a specific ship, providing a stable and unchanging ideal benchmark for all subsequent parallel operations and game theory analyses. The ideal parameter acquisition unit within the module systematically collects and solidifies the core dataset constituting the ideal ship performance profile. This unit performs a one-time data import operation, digitizing and storing a series of ideal physical parameters from tank test reports provided by the ship design institute, open-water test data provided by the propeller manufacturer, and bench test reports provided by the main engine manufacturer. These key parameters mainly include the smooth hull resistance curve, existing as discrete data points or fitted curves, describing the pure resistance of the new ship at different speeds; the new propeller open-water characteristic curve, defining the propeller's thrust, torque, and efficiency characteristics in uniform flow; and the main engine standard efficiency spectrum, calibrating the fuel consumption rate of the main engine at different loads and speeds. The simulation building unit integrates the discrete ideal physical parameters into a mathematical model with dynamic simulation capabilities through the principles of fluid mechanics and ship kinematics. This unit is based on classical ship motion equations, such as three- or six-degree-of-freedom maneuvering models, and embeds parameters provided by the ideal parameter acquisition unit as core coefficients of the model. The core function of this model is to calculate the theoretical propulsion power required to maintain the input speed, and the calculation process is characterized by the following formula: ; in, This represents the theoretically optimal power calculated by the high-fidelity virtual model, in kilowatts. It is the ship's total theoretical resistance, and it is the speed. The function, whose value is obtained by interpolation or function call based on the resistance curve of the smooth hull. It is the ship's speed over water. This is the comprehensive propulsion efficiency, a dimensionless coefficient calculated from ideal parameters such as the new propeller's open-water characteristic curve, hull efficiency, and relative rotational efficiency. Through this construction process, a high-fidelity virtual model is generated that, upon receiving navigation control commands, can accurately simulate and calculate the ship's theoretically expected energy consumption under ideal conditions and its theoretically optimal power at a specific speed, laying the foundation for subsequent parallel operation and game theory processing.

[0046] For example, when a certain type of bulk carrier is deployed for the first time, the model building module constructs an ideal benchmark model. The ideal parameter acquisition unit first digitally stores the smooth hull resistance curve from the tank test report provided by the ship design institute, and obtains the open water characteristic curve of the new propeller and the standard efficiency spectrum of the main engine. Subsequently, the simulation building unit, based on the three-degree-of-freedom maneuvering motion model, embeds the above parameters into the ship's kinematics equations, assuming the actual ship's current speed in the water. The speed is 15 knots, approximately 7.72 meters per second. The total theoretical resistance at this speed is obtained by interpolation using the drag curve of a smooth hull. The overall propulsion efficiency was calculated from parameters such as the propeller open-water characteristic curve, with a value of 850 kN. The value is 0.72, which is then substituted into the formula. Calculations were performed to obtain the theoretically optimal power required to maintain this speed. The high-fidelity virtual model constructed from this kilowatt accurately simulates the energy consumption benchmark of the ship under ideal conditions, providing a stable reference for energy efficiency competition in subsequent parallel operations.

[0047] In one possible implementation, combining Figure 2 The decision output module includes: The state quantification unit is used to quantify the evolved core physical parameters in the dynamic digital image into a ship physical condition deterioration certificate. The diagnostic generation unit is used to analyze the contribution ratio of capability residuals and strategy residuals to the total energy consumption deviation and generate an energy efficiency performance diagnostic report. The maintenance triggering unit is used to automatically generate a predictive maintenance work order that includes a certificate of physical condition deterioration of the ship and a diagnostic report of its operational performance when the long-term trend value of the capability residual exceeds a preset maintenance threshold.

[0048] In some implementations, the decision output module serves as the final execution and human-computer interaction layer, transforming the results of analysis and optimization into clearly guiding assessment reports and automated instructions, thus closing the complete loop from data analysis to actual decision-making. The state quantification unit within this module transforms abstract parameter changes in a dynamic digital image into intuitively measurable ship health status indicators. This unit periodically extracts core physical parameters corrected by the model evolution engine, such as hull resistance and propeller efficiency parameters, and compares them with the ideal physical parameters stored during initial construction. By calculating the relative rate of change of these parameters, a ship physical condition deterioration certificate is generated. This certificate is not a paper document in engineering but a structured data report, its core being quantified deterioration indicators, such as a percentage increase in hull resistance compared to a new ship condition, or a percentage decrease in propulsion efficiency. The diagnostic generation unit aims to accurately analyze the causes of total energy consumption deviations, providing clear attribution directions for subsequent improvement measures. This unit obtains the average value of the capability residual and strategy residual over a voyage or specific time period, such as 24 hours, from the game processing module. The contribution ratio of both to the total deviation is calculated using the following formula: ; in, It is the contribution ratio of the ability residual. and These are the average values ​​of the capability residual and the strategy residual over the given time period, both measured in kilowatts (kW), and can therefore be directly added. This unit generates an energy efficiency performance diagnosis report based on the calculated contribution ratio, clearly indicating the main reasons for the current poor energy efficiency, such as "70% of the total energy consumption deviation in the past 24 hours comes from the strategy residual, and it is recommended to optimize speed and heading control," or "85% of the total energy consumption deviation comes from the capability residual, indicating severe deterioration of the ship's physical condition." The maintenance triggering unit enables condition-based predictive maintenance, transforming passive repair into proactive maintenance. This unit continuously monitors the long-term trend value of the capability residual, which is typically obtained through time-series analysis of capability residual data from the past 30 to 90 days, such as linear regression or exponential smoothing. When this trend value exceeds a preset maintenance threshold, the maintenance process is automatically triggered. This process is represented by the conditional judgment of the following formula: ; in, It is the long-term trend value of the capacity residual, measured in kilowatts, reflecting the continuous power loss caused by physical degradation. This is the maintenance threshold, a key engineering parameter set based on economic analysis. It typically corresponds to the point at which the additional fuel cost will soon exceed the cost of a single hull cleaning or propeller polishing, and its value may range from hundreds to thousands of kilowatts. Once the condition is met, the unit immediately and automatically generates a predictive maintenance work order. This work order embeds the ship's physical condition deterioration certificate generated by the condition quantification unit and the energy efficiency performance diagnosis report generated by the diagnosis generation unit, and sends it to the shore-based management department's maintenance planning system to provide a basis for decision-making.

[0049] For example, the decision output module evaluates and analyzes the energy efficiency performance of a certain type of ocean-going cargo ship over the past 24 hours and longer periods. The state quantification unit extracts the current hull resistance parameters from the dynamic digital image and compares them with the ideal parameters, quantifying and generating a certificate of physical condition deterioration of the ship, showing an 8.5% increase in additional hull resistance. Subsequently, the diagnosis generation unit obtains the average value of the capability residuals over the past 24 hours. For 350 kilowatts, the average value of the strategy residuals 150 kilowatts, substitute into the formula The contribution ratio of capability residuals was calculated. Based on this, an energy efficiency performance diagnosis report is generated, indicating that the main reason for the current poor energy efficiency is the deterioration of the ship's physical condition; finally, the maintenance triggering unit obtains the long-term trend value of the capability residual by performing exponential smoothing analysis on the data from the past 60 days. The power is 420 kilowatts, if the preset maintenance threshold is... 400 kilowatts, substitute into the formula The determination is based on 420 kilowatts. At 400 kilowatts, the conditions are met and the predictive maintenance process is automatically triggered, generating a predictive maintenance work order that includes the aforementioned degradation certificate and diagnostic report, and sending it to the maintenance planning system.

[0050] In one possible implementation, combining Figure 2 The system also includes a digital resume module: The sequence recording unit is used to continuously record the values ​​of the core physical parameters of the model evolution engine after each iteration adjustment and the corresponding timestamp, forming a parameter evolution sequence; The associated construction unit is used to associate the parameter evolution sequence with the navigation log and maintenance records of the same period to construct a timeline of ship performance degradation; The trend prediction unit is used to generate predictions of future energy efficiency status based on the performance degradation timeline and using a trend prediction model.

[0051] In some implementations, the digital history module provides data memory and forward-looking capabilities spanning the entire lifecycle of a ship. It systematically records and correlates the evolution of the ship's physical performance with significant events, constructing a timeline of ship performance degradation for long-term analysis and future prediction. The sequence recording unit within the module provides the foundational time-series data for creating a dynamic digital mirror of the core parameter evolution. This unit is passively triggered each time the model evolution engine module performs an iterative correction operation. Once the core physical parameters, namely hull drag and propeller efficiency, are updated, this unit immediately captures the updated values ​​and appends a timestamp accurate to the second. By continuously performing this operation, this unit generates and maintains two parallel parameter evolution sequences in the database, fully recording the minute quantitative evolution of the ship's physical state over time. The correlation construction unit imbues pure parameter changes with rich engineering context and physical meaning, transforming data into information. This unit executes periodically, for example daily, or is triggered when a significant step or inflection point is detected in the parameter evolution sequence. It aligns and correlates the parameter evolution sequence temporally with concurrent electronic navigation logs and digital maintenance records obtained from the ship's integrated management system. For example, if the hull resistance parameter drops significantly at a certain point in time, this unit will automatically retrieve maintenance records near that point. If a matching "hull cleaning" or "dock repair" record is found, the event is marked at the corresponding point in the parameter sequence. In this way, isolated parameter data points are linked together to form a ship performance degradation timeline rich in contextual information. The trend prediction unit transforms historical data into actionable insights for the future, supporting longer-term planning. This unit operates based on the ship performance degradation timeline generated by the association building unit. One or more trend prediction models, such as the autoregressive integral moving average model ARIMA or a deep learning-based long short-term memory network LSTM, are used to train and extrapolate the parameter evolution sequence. The prediction process is represented by the following formula: ; in, These are predicted values ​​for future core physical parameters. This represents the selected trend prediction model function. It is a pure degradation history parameter sequence extracted from the ship performance degradation timeline, after removing the step effects caused by maintenance events. This is the predicted time span, a parameter that can be set by the user, typically 6 to 24 months into the future. The unit's output estimates, such as "It is expected that within the next 9 months, without intervention, the additional drag on the hull will increase by another 5%, resulting in an average increase of 1.2 tons of additional fuel consumption per day," provide shipping companies with a reliable quantitative basis for developing long-term maintenance budgets and dry-docking plans.

[0052] For example, the digital history module records and predicts the long-term performance evolution of a certain type of Panamax bulk carrier. The sequence recording unit is passively triggered when the model evolution engine updates the core physical parameters, capturing in real time the hull resistance parameter from 1.055 to 1.058 and attaching a precise timestamp to generate a continuously evolving parameter evolution sequence. Subsequently, when the association construction unit detects a significant step drop in the resistance parameter in the parameter evolution sequence, it automatically associates it with the "dock entry for painting" record in the electronic navigation log of the same period, connecting isolated data points into a timeline of ship performance degradation with an engineering background. Finally, the trend prediction unit extracts the pure deterioration historical parameter sequence after removing maintenance event interference. Select the LSTM prediction model function And set the prediction time span Substitute 12 months into the formula. Extrapolation calculations were performed to obtain the predicted hull resistance parameters for the next 12 months. The value is 1.115, and based on this, it is estimated that without intervention, it will lead to an additional 1.5 tons of fuel consumption per day, providing a quantitative basis for shipping companies to formulate annual dry-docking plans.

[0053] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0054] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the disclosure of the specification and practices. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A ship energy efficiency optimization management and evaluation system based on big data analysis, characterized in that, The system includes: The command synchronization module is used to acquire the navigation control commands of the actual ship and synchronously send the navigation control commands to a preset high-fidelity virtual model for simulating the ideal physical characteristics of the ship, drive the high-fidelity virtual model to run in parallel with the actual ship, and generate parallel operation data. The game theory processing module is used to perform two-way expected game processing based on the parallel operation data. It calculates the capability residual by determining the first difference between the actual energy consumption of the actual ship and the theoretical expected energy consumption output by the high-fidelity virtual model, and calculates the second difference between the actual power of the actual ship's main engine and the theoretical optimal power derived from the high-fidelity virtual model, thus obtaining the strategy residual. The game theory processing module includes a forward comparison unit, a reverse comparison unit, and a structured storage unit. The forward comparison unit extracts the theoretical expected energy consumption and actual energy consumption from the parallel operation data, performs forward comparison processing, and calculates and outputs the capability residual. The reverse comparison unit extracts the actual water speed reached by the actual ship from the parallel operation data, drives the high-fidelity virtual model to derive the theoretical optimal power based on the water speed, and simultaneously obtains the actual power of the actual ship's main engine, performs reverse comparison processing, and calculates and outputs the strategy residual. The structured storage unit establishes the association between the capability residual and the strategy residual and the current navigation condition label, forming structured game difference data. The model evolution engine module is used to iteratively adjust the core physical parameters in the high-fidelity virtual model based on the capability residuals, so that the high-fidelity virtual model evolves into a dynamic digital image representing the current physical state of the actual ship. The model evolution engine module includes: a pattern recognition unit, a parameter mapping unit, and an iterative correction unit. The pattern recognition unit is used to perform pattern recognition processing on the capability residuals, distinguishing between a first type of residual pattern caused by changes in hull resistance and a second type of residual pattern caused by changes in propulsion efficiency. The parameter mapping unit is used to map the first type of residual pattern to an adjustment amount for the hull resistance parameter in the high-fidelity virtual model, and to map the second type of residual pattern to an adjustment amount for the propeller efficiency parameter in the high-fidelity virtual model. The iterative correction unit is used to iteratively correct the hull resistance parameter and the propeller efficiency parameter based on the adjustment amounts of the hull resistance parameter and the propeller efficiency parameter, completing the autonomous evolution of the high-fidelity virtual model into a dynamic digital image. The optimization and verification loop module is used to perform simulation optimization using the evolved dynamic digital image, generating an optimized strategy package with the goal of reducing the strategy residual, and performing closed-loop verification of the optimized strategy package by executing the two-way expected game processing to obtain a verified optimized strategy. The optimization and verification loop module includes: a simulation optimization unit, a strategy verification unit, and a threshold determination unit. The simulation optimization unit is used to perform combined optimization of airspeed setting, heading, and wing deflection angle in the dynamic digital image, searching for parameter combinations that can optimize the predicted strategy residual to form an initial optimized strategy package. The strategy verification unit is used to use the initial optimized strategy package as a new navigation control command to drive the execution of the two-way expected game processing to obtain a strategy residual for verification. The threshold determination unit is used to determine whether the strategy residual for verification meets a preset optimization effectiveness threshold. If yes, the initial optimized strategy package is marked as a verified optimized strategy; otherwise, the initial optimized strategy package is discarded and a new round of simulation optimization is triggered. The decision output module is used to generate a comprehensive evaluation report and trigger corresponding maintenance or execution instructions based on the parameters, capability residuals, strategy residuals, and verified optimization strategies of the dynamic digital image.

2. The big data analytics based ship energy efficiency optimization management and evaluation system according to claim 1, characterized in that, The pattern recognition unit includes: The linear identification subunit is used to calculate the first derivative of the capability residual with time series to identify a linear growth trend, and to calculate the correlation coefficient between the capability residual and environmental parameters to eliminate interference from environmental factors, thereby identifying the first type of residual pattern. The correlation analysis subunit is used to subtract the first type of residual pattern from the capability residual to obtain the residual fluctuation component, and to calculate the correlation strength between the residual fluctuation component and the host operating parameters, thereby identifying the second type of residual pattern.

3. The ship energy efficiency optimization management and evaluation system based on big data analysis according to claim 1, characterized in that, The simulation optimization unit includes: The constraint acquisition subunit is used to acquire the expected arrival time constraints and the carbon emission intensity indicators required by regulations in the voyage plan; The boundary processing subunit is used to treat the expected arrival time constraint as a hard boundary condition for combined optimization and the carbon emission intensity index constraint as an optimization objective in multi-objective optimization calculation. The solution execution subunit is used to solve the multi-objective optimization calculation under the premise of satisfying the hard boundary conditions, and generate an initial optimization strategy package.

4. The ship energy efficiency optimization management and evaluation system based on big data analysis according to claim 1, characterized in that, The system also includes a model building module: The ideal parameter acquisition unit is used to acquire ideal physical parameters characterizing the performance of the ship during the design phase. The ideal physical parameters include the resistance curve of the smooth hull, the open water characteristic curve of the new propeller, and the standard efficiency spectrum of the main engine. The simulation construction unit is used to construct a high-fidelity virtual model based on the ideal physical parameters and the ship's motion equations, capable of simulating and calculating the theoretical expected energy consumption and the theoretical optimal power.

5. The ship energy efficiency optimization management and evaluation system based on big data analysis according to claim 1, characterized in that, The decision output module includes: The state quantization unit is used to quantify the evolved core physical parameters in the dynamic digital image into a ship physical condition deterioration certificate. The diagnostic generation unit is used to analyze the contribution ratio of the capability residual and strategy residual to the total energy consumption deviation and generate an energy efficiency performance diagnostic report. The maintenance triggering unit is used to automatically generate a predictive maintenance work order containing the ship's physical condition deterioration certificate and the capability performance diagnosis report when the long-term trend value of the capability residual exceeds a preset maintenance threshold.

6. The ship energy efficiency optimization management and evaluation system based on big data analysis according to claim 1, characterized in that, The system also includes a digital resume module: The sequence recording unit is used to continuously record the values ​​of the core physical parameters of the model evolution engine after each iteration adjustment and the corresponding timestamps, forming a parameter evolution sequence; The association construction unit is used to associate the parameter evolution sequence with the navigation log and maintenance records of the same period to construct a timeline of ship performance degradation; The trend prediction unit is used to generate a prediction of future energy efficiency status based on the performance degradation timeline using a trend prediction model.

7. A method for ship energy efficiency optimization management and evaluation based on big data analysis, characterized in that, The method is used in the ship energy efficiency optimization management and evaluation system based on big data analysis as described in any one of claims 1-6, and the method includes: Obtain navigation control commands from the actual ship and synchronously send the navigation control commands to a preset high-fidelity virtual model used to simulate the ideal physical characteristics of the ship, drive the high-fidelity virtual model to run in parallel with the actual ship, and generate parallel operation data; Based on the parallel operation data, a two-way expected game process is performed. The capability residual is obtained by calculating the first difference between the actual energy consumption of the actual ship and the theoretical expected energy consumption output by the high-fidelity virtual model, and the strategy residual is obtained by calculating the second difference between the actual power of the actual ship's main engine and the theoretical optimal power derived by the high-fidelity virtual model. This includes: extracting the theoretical expected energy consumption and actual energy consumption from the parallel operation data, performing a forward comparison, calculating and outputting the capability residual; extracting the actual water speed reached by the actual ship from the parallel operation data, driving the high-fidelity virtual model to derive the theoretical optimal power based on the water speed, and simultaneously obtaining the actual power of the actual ship's main engine, performing a reverse comparison, calculating and outputting the strategy residual; establishing the correlation between the capability residual and the strategy residual and the current navigation condition label, forming structured game difference data. Based on the capability residuals, the core physical parameters in the high-fidelity virtual model are iteratively adjusted to evolve the high-fidelity virtual model into a dynamic digital image representing the current physical state of the actual ship. This includes: performing pattern recognition processing on the capability residuals to distinguish between a first type of residual pattern caused by changes in hull resistance and a second type of residual pattern caused by changes in propulsion efficiency; mapping the first type of residual pattern to an adjustment amount for the hull resistance parameter in the high-fidelity virtual model, and mapping the second type of residual pattern to an adjustment amount for the propeller efficiency parameter in the high-fidelity virtual model; iteratively correcting the hull resistance parameter and propeller efficiency parameter based on the adjustment amounts of the hull resistance parameter and propeller efficiency parameter, thus completing the autonomous evolution of the high-fidelity virtual model into a dynamic digital image. The evolved dynamic digital mirror is used for simulation optimization to generate an optimized strategy package with the goal of reducing the strategy residual. The two-way expected game process is then executed to perform closed-loop verification of the optimized strategy package, resulting in a verified optimized strategy. This process includes: optimizing the speed setting, heading, and wing deflection angle in the dynamic digital mirror to search for parameter combinations that optimize the predicted strategy residual, forming an initial optimized strategy package; using the initial optimized strategy package as a new navigation control command to drive the execution of the two-way expected game process, obtaining a strategy residual for verification; and determining whether the strategy residual for verification meets a preset optimization effectiveness threshold. If yes, the initial optimized strategy package is marked as a verified optimized strategy; otherwise, the initial optimized strategy package is discarded and a new round of simulation optimization is triggered. Based on the parameters, capability residuals, strategy residuals, and verified optimization strategies of the dynamic digital image, a comprehensive evaluation report is generated and corresponding maintenance or execution instructions are triggered.

Images