A Method for Energy Management and Efficiency Optimization of Ship Shaft-Driven Power Generation Based on Multi-Operating Condition Adaptation

Abstract

This invention relates to the field of ship electrical system control technology, and particularly to a method for ship shaft-driven power generation energy management and efficiency optimization based on multi-condition adaptation. The method includes: a data acquisition and processing step: real-time acquisition and preprocessing of multi-source heterogeneous operating parameters that comprehensively characterize the ship's propulsion system, electrical system, motion state, and external environment; a dynamic condition identification step; a dynamic optimization model construction step; an online optimization and control step; and an adaptive learning step: periodically incrementally training the time-series deep learning model based on new data generated during system operation to update its network parameters. Adaptive learning enables the system to continuously optimize the accuracy of condition identification as the ship operates over time and the external environment changes, improving the system's long-term stability and adaptability.

Description

Technical Field

[0001] This invention relates to the field of ship power system control technology, and in particular to a method for ship shaft-driven power generation energy management and efficiency optimization based on multi-condition adaptation. Background Technology

[0002] Marine shaft generators (SGs) are important energy-saving devices that utilize the surplus power of the main engine to generate electricity. However, existing energy management methods for shaft generator systems have significant shortcomings:

[0003] The current technology suffers from crude operating condition identification and poor adaptability: It typically relies on only a single or a few parameters, such as main engine speed or power, to roughly determine a ship's operating condition (e.g., "constant speed navigation" or "maneuvering navigation"), lacking a refined and multi-dimensional identification of the ship's actual and complex operating state. For example, even at the same main engine rated speed, a ship may be operating under drastically different conditions, such as headwinds, tailwinds, light loads, or heavy loads, resulting in significant differences in its main engine surplus power and grid demand fluctuations. This crude identification leads to inaccurate energy management strategies, either being overly conservative and wasting energy or overly aggressive and compromising main engine safety.

[0004] Energy management is static and lacks foresight and coordination: Most existing solutions adopt energy allocation strategies based on fixed rules or lookup tables, and their control parameters (such as power allocation ratios and bus voltage thresholds) are usually preset and static. This method cannot respond to dynamically changing surplus power and randomly fluctuating electrical loads during ship navigation, let alone cope with upcoming operational condition transitions (such as upcoming turning or avoidance maneuvers). There is a lack of coordinated optimization among energy management units, shaft generators, auxiliary generators, and energy storage devices, often resulting in a piecemeal approach and low overall system efficiency.

[0005] Efficiency optimization is often localized and lacks a systematic approach: Existing optimization methods often focus only on the instantaneous efficiency of the shaft-driven generator itself or a single point on the main engine, neglecting the fact that the shaft-driven power generation system is a coupled system involving multiple energy conversion stages, including mechanical, electrical, and thermal processes. For example, blindly pursuing high power output from the shaft-driven generator may cause the main engine to operate in an inefficient zone, thereby increasing fuel consumption; or frequent start-ups and shutdowns of auxiliary engines to stabilize grid frequency may lead to a decrease in the average load rate of the auxiliary engines, resulting in deterioration of their efficiency. This approach of localized optimization rather than global optimization makes it difficult to achieve a true improvement in the overall ship energy efficiency.

[0006] Therefore, there is an urgent need for a method that can accurately perceive the overall operating conditions of a ship and, based on this, perform dynamic, forward-looking, and collaborative global energy management and efficiency optimization. Summary of the Invention

[0007] To achieve the above objectives, this invention provides a method for energy management and efficiency optimization of ship shaft-driven power generation based on multi-condition adaptation, characterized by comprising the following steps:

[0008] Data acquisition and processing steps: Real-time acquisition and preprocessing of multi-source heterogeneous operating parameters, including main engine parameter group, ship motion and environment parameter group, power system parameter group, and navigation command parameter group;

[0009] Operating condition dynamic identification steps: The processed main engine parameter group, ship motion and environment parameter group, power system parameter group, and navigation command parameter group are input into a pre-trained time series deep learning model to calculate the operating condition identification result. The operating condition identification result is used to characterize the confidence level of the ship in various predefined refined operating conditions at the current moment. The operating condition identification result includes the dominant operating condition. The refined operating condition is a typical operating state that is distinguished from the traditional dichotomy method and is divided according to the clustering of multi-dimensional operating parameter combinations.

[0010] The optimization model dynamic construction steps are as follows: Based on the dominant operating condition in the operating condition identification results, a global efficiency optimization model is selected and constructed. The optimization objective of the global efficiency optimization model is to maximize the overall efficiency of the entire path from the fuel consumption of the main engine to the power output of the grid. The decision variables of the global efficiency optimization model include the power commands of the shaft generator, the auxiliary generator and the energy storage system. The constraint coefficients of the global efficiency optimization model are adaptively adjusted according to the dominant operating condition.

[0011] Online optimization and control steps: The constructed global efficiency optimization model is solved using an online optimization algorithm, and the optimal decision variable values ​​obtained are used as control commands and synchronously sent to the actuators of the shaft generator, auxiliary generator and energy storage system.

[0012] Adaptive learning steps: Based on new data generated during system operation, the time-series deep learning model is periodically incrementally trained to update its network parameters.

[0013] Preferably, the temporal deep learning model is a classification model that combines a long short-term memory network with an attention mechanism;

[0014] The pre-training includes: training the model using a historical dataset containing time series samples of the multi-source heterogeneous operating parameters and their corresponding manually labeled operating conditions. During the training process, the cross-entropy loss between the model's prediction results and the real labels is minimized through the backpropagation algorithm, thereby learning the nonlinear mapping relationship between complex multidimensional input sequences and refined operating condition classification.

[0015] The operating condition identification result is a multi-dimensional vector. The value of each element in the vector represents the confidence probability that the model determines the current state belongs to a certain predefined refined operating condition. The dominant operating condition is the operating condition category with the highest confidence probability value.

[0016] Preferably, the predefined fine chemical conditions include at least the following categories:

[0017] Rated speed with wind and waves during stable navigation: The main engine speed is stable near the rated value, the ship speed is high and fluctuates little, the pitch and roll angles are low, the wind and waves come from the stern, and the power grid load is stable.

[0018] Rated speed against wind and waves under heavy load navigation conditions: The main engine speed is stable at the rated value, the ship speed is lower than expected, the pitch angle amplitude is large and the period is fixed, the wind and waves come from the bow direction, and the main engine output torque and fuel injection quantity are close to the upper limit.

[0019] Deceleration maneuvering and large rudder angle turning conditions: These conditions are characterized by the main engine speed being in the process of decreasing or below the rated value, the rudder angle command setting value continuously deviating from zero, the ship's speed and course continuously changing, and the power grid load possibly experiencing large fluctuations.

[0020] The set of categories for the refined chemical conditions is summarized and defined by applying an unsupervised clustering algorithm to a large amount of historical data and combining it with domain expert knowledge.

[0021] Preferably, in the dynamic construction step of the optimization model, the objective function of the global efficiency optimization model is the mathematical representation of the overall efficiency of the entire path. The calculation of this efficiency needs to comprehensively consider the fuel consumption rate of the host at the current load point, the power generation efficiency of the shaft generator, the power generation efficiency of each auxiliary generator in operation at its specific load point, and the charging and discharging efficiency of the energy storage system.

[0022] The specific form of the objective function is adjusted according to the optimization focus implied by the dominant operating condition.

[0023] Preferably, the adaptive adjustment of the constraint coefficients refers to:

[0024] The upper limit of the safe operating range of the total host load in the host load constraint is set higher when the host has sufficient spare power to make full use of energy, and lower when the host has limited spare power to reserve more power margin for the propulsion system.

[0025] The high-efficiency operating range of a single auxiliary generator in the auxiliary generator operation constraints is dynamically calculated based on the number of auxiliary generators currently in operation and the total grid load, to ensure that the load rate of each operating auxiliary generator falls as close as possible to the high-efficiency segment of its efficiency characteristic curve.

[0026] The expected value of the target state of charge in the operation constraints of the energy storage system is set according to the prediction of the operating condition trend in the future. If it is predicted that the system will enter a high load condition, the expected value is increased to store energy. If it is predicted that the system will enter a low load condition, the expected value is decreased to prepare to absorb excess energy.

[0027] Preferably, in the online optimization and control step, the online optimization algorithm used is one of linear programming, quadratic programming, or genetic algorithm. The algorithm selection strategy is as follows: when the global efficiency optimization model is determined to be a convex optimization problem, linear programming or quadratic programming algorithms with fast computation speed are preferred for accurate solution; when the model contains non-convex or discrete variables, genetic algorithm is used for heuristic search to obtain a suboptimal solution that meets the requirements of engineering applications within an acceptable time.

[0028] Preferably, the synchronous issuance of control commands means that the target active power command of the shaft generator, the number of auxiliary generators to be switched and the target power allocation command, and the target charging and discharging power command of the energy storage system, calculated within a control cycle, are sent to their respective execution controllers through the ship control network within the same communication cycle. Each controller is required to start executing the new command at the next unified control time starting point, thereby ensuring coordinated operation between multiple energy subsystems and avoiding power oscillations or system instability caused by asynchronous commands.

[0029] Preferably, in the adaptive learning step, the specific process of periodic incremental training is as follows: the system continuously stores the actual data in operation, including the input multi-source heterogeneous operating parameters, the operating condition identification results output by the model, the issued control commands, and the changes in system state after execution, as new training samples in the buffer; every fixed time period or when the number of new samples reaches a preset scale, a portion of samples is extracted from the buffer to perform a round of incremental training on the time-series deep learning model; a small learning rate is used during training to fine-tune the network parameters, so that the model can slowly adapt to the drift in operating characteristics caused by ship fouling and main engine performance degradation.

[0030] Preferably, in the data acquisition and processing steps,

[0031] The collected multi-source heterogeneous operating parameters specifically include:

[0032] Main unit parameter group: The main unit fuel engine shaft speed and main unit output torque are collected by sensors installed on the main unit, and the main unit fuel injection quantity, main unit scavenging pressure and main unit exhaust temperature are read by the main unit control system;

[0033] Ship motion and environmental parameters group: The ship's speed above ground is obtained through a global positioning system receiver; the pitch and roll angles are obtained through an attitude sensor installed near the ship's center of gravity; the relative wind speed and direction are obtained through a wind speed and direction sensor installed at the top of the mast; and the wave height is estimated by analyzing the ship's motion response and combining it with the model.

[0034] Power system parameter group: The active and reactive power of the shaft generator are collected by the current transformer and voltage transformer installed on the output circuit of the shaft generator, the total active load of the power grid and the bus voltage are collected by the current transformer and voltage transformer installed on the main switchboard bus, and the current state of charge of the energy storage system is read through the communication interface of the battery management system built into the energy storage system.

[0035] Navigation command parameter group: Reads the current main engine speed command setting value and rudder angle command setting value from the bridge control system and steering gear control system via the ship's local area network;

[0036] The preprocessing includes: First, using a moving average filtering method to smooth all parameters except the rudder angle command to suppress high-frequency noise. The time length of the moving average window is preset according to the physical characteristics and rate of change of the parameters themselves. Second, using a limiting filtering method to process all parameters to eliminate physically impossible outliers caused by instantaneous sensor interference. The limiting threshold is determined based on the historical maximum and minimum values ​​of each parameter with a certain margin. Finally, all parameters are labeled with a uniform time stamp and aligned to ensure that data at the same moment correspond to the same sampling period.

[0037] Preferably, the method forms a closed-loop control system, wherein the data acquisition and processing step, the dynamic identification step of operating conditions, the dynamic construction step of the optimization model, and the online optimization and control step are executed periodically, and the cycle period is determined according to the speed of the dynamic response of the ship's power system.

[0038] The beneficial effects of this invention are:

[0039] 1. This invention proposes a ship operating condition identification method based on multi-dimensional perception and dynamic adaptation. By comprehensively considering multiple real-time monitoring parameters such as main engine speed, power, speed, sea state, and ship load, it accurately identifies the specific operating condition of the ship. This refined operating condition identification method ensures the accurate adaptation of energy management strategies under complex operating conditions, avoiding energy waste and main engine safety risks in traditional methods, thereby improving the adaptability and accuracy of energy management.

[0040] 2. This invention introduces a dynamic energy allocation strategy based on a predictive model, enabling real-time response to changes in surplus power and load fluctuations during ship navigation. Particularly for upcoming operational condition transitions, the invention's forward-looking energy management method adjusts energy allocation based on predictive information, thereby optimizing the overall system's energy efficiency. Furthermore, this invention employs a collaborative scheduling mechanism based on a global optimization algorithm, significantly improving the coordination between the energy management unit, shaft generator, auxiliary generator, and energy storage device. This avoids local optimization problems associated with individual devices, thereby enhancing the overall system efficiency.

[0041] 3. This invention proposes a global optimization method that integrates the coupling relationships between mechanical, electronic, and thermal systems to comprehensively optimize the efficiency of a ship's shaft-driven power generation system. Through multi-stage integrated and coordinated optimization, it avoids the localized optimization of a single point, such as the shaft-driven generator or the main engine, as seen in traditional methods. This method can improve the power output of the shaft-driven generator while preventing the main engine from operating in an inefficient range, reducing fuel consumption, and maintaining the optimal load rate of auxiliary generators by optimizing their start-stop frequency, thereby improving the overall system's energy utilization efficiency and achieving a comprehensive improvement in the ship's overall energy efficiency. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of the steps of the method of the present invention;

[0044] Figure 2 This is a flowchart illustrating the adaptive adjustment steps of the constraint coefficients in the method of the present invention. Detailed Implementation

[0045] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0046] Please see Figures 1-2This invention provides a method for energy management and efficiency optimization of ship shaft-driven power generation based on multi-condition adaptation. It collects ship operating data in real time and preprocesses this data. The collected parameters include ship propulsion system data (such as main engine speed and power output), power system data (such as grid voltage, current, and power demand), ship motion status (such as speed, heading, and hull attitude), and external environmental data (such as sea state, wind speed, and temperature). These parameters come from various sensors on the ship, such as speed sensors, temperature sensors, and pressure sensors. To ensure data accuracy and real-time performance, the collected data needs to undergo preprocessing operations such as filtering, noise reduction, and standardization to ensure that subsequent analysis effectively reflects the actual operating conditions of the ship.

[0047] The data acquisition and processing steps ensure the acquisition of multi-dimensional, real-time operational data from the vessel, providing a foundation for accurate monitoring of its operating conditions. This data processing flow provides reliable and accurate input for subsequent condition identification and optimization models.

[0048] Preprocessed multi-source heterogeneous data is input into a pre-trained temporal deep learning model. This model learns from the ship's historical operational data to capture the ship's operational characteristics under different conditions. Through this model, the system can calculate the confidence levels of various predefined operating conditions the ship is currently in and identify these conditions. Operating conditions are not solely determined by traditional speed or RPM, but are refined through a combination of parameters from multiple dimensions, enabling the identification of complex conditions that differ from traditional binary classifications, such as "navigating against waves" and "navigating under heavy load."

[0049] By introducing deep learning technology, operating condition identification becomes more accurate, enabling real-time reflection of the ship's complex operating status. Compared to the coarse judgment of traditional methods, the dynamic identification method provided by this invention can fully adapt to changing operating conditions, avoiding the failure of energy management strategies due to inaccurate operating condition identification.

[0050] Based on the dominant operating conditions identified in the operating condition identification results, the system dynamically selects and constructs a global efficiency optimization model. The goal of this model is to maximize the overall efficiency from main engine fuel consumption to grid power output by adjusting the power commands of various ship energy systems (such as shaft generators, auxiliary generators, and energy storage systems). The constraints of the optimization model are dynamically adjusted according to the currently identified dominant operating conditions. For example, when the ship is lightly loaded, the model may prioritize fuel efficiency, while under heavy load, greater consideration needs to be given to the stability of the power supply.

[0051] The dynamically constructed optimization model can flexibly adjust the energy allocation strategy according to real-time operating condition changes, making the operation of the entire system more in line with the actual needs of the ship. Through a global optimization method, this invention can comprehensively consider the coordinated operation of different devices, improve the overall energy efficiency of the system, and avoid energy waste or unstable power output caused by static strategies.

[0052] An online optimization algorithm is used to solve the global efficiency optimization model in real time, and the obtained optimal decision variables (such as the power of each generator and the power of the energy storage system) are distributed to the corresponding actuators. Through real-time optimization, the system can dynamically respond to various changes during the ship's navigation, such as load fluctuations and changes in power grid frequency, ensuring the coordinated operation of various energy devices on the ship.

[0053] The real-time nature and flexibility of online optimization ensure that ships can obtain optimal energy allocation under different navigation conditions, avoiding the lag and maladaptation problems caused by traditional static methods, and enhancing the system's responsiveness and stability.

[0054] During ship operation, new data is constantly generated. The system will periodically and incrementally train the time-series deep learning model based on this new data, updating the network parameters so that the model can adapt to dynamic changes during long-term operation. This step can continuously optimize the operating condition identification capability and ensure the accuracy and reliability of the model after long-term operation.

[0055] Adaptive learning enables the system to continuously optimize the accuracy of operating condition identification as the ship operates over time and the external environment changes, thereby improving the system's long-term stability and adaptability. In this way, the model can more accurately identify complex and changing operating conditions, continuously improving overall energy efficiency.

[0056] By employing precise data acquisition and processing, dynamic operating condition identification, construction of a global optimization model, real-time online optimization and control, and an adaptive learning mechanism, the energy efficiency and stability of the ship's shaft-driven power generation system are comprehensively improved. Compared with existing technologies, this method offers higher adaptability, flexibility, and efficiency optimization capabilities, significantly enhancing the overall energy management level of ships.

[0057] In one possible implementation, this method involves a data acquisition phase where various sensors installed at different locations on the ship comprehensively acquire multi-dimensional operational parameters. These parameters are categorized into four main types:

[0058] Sensors installed on the main unit collect core data in real time, such as the main unit's engine shaft speed and output torque. Simultaneously, the main unit's control system reads operating status data such as fuel injection quantity, scavenging pressure, and exhaust temperature. This data helps to comprehensively understand the main unit's operating status, ensuring that its power output remains within a reasonable range and preventing excessive fuel consumption or mechanical overload.

[0059] The ship's ground speed is obtained through the Global Positioning System (GPS), while attitude sensors near the ship's center of gravity measure pitch and roll angles to understand changes in the ship's attitude. Wind speed and direction sensors at the masthead provide wind speed and direction information to help assess the impact of wind forces during navigation. Furthermore, by combining the ship's motion response with models, wave height can be estimated. This data is crucial for analyzing the ship's power requirements under different navigation conditions.

[0060] In power systems, current transformers and voltage transformers collect data on the active and reactive power of shaft-driven generators, as well as the active load and bus voltage of the power grid. This power data reflects the system's load status and helps adjust the output of generators and energy storage systems. Furthermore, the state of charge of energy storage devices is read through the communication interface of the energy storage system's built-in battery management system, thereby assessing the system's energy storage capacity and optimizing energy distribution.

[0061] Navigation commands are primarily obtained through the ship's local area network, including main engine speed commands and rudder angle commands. By adjusting these command values ​​in real time, the system can effectively control the ship's speed and direction, ensuring stable operation under predetermined conditions.

[0062] Collected heterogeneous data from multiple sources often contains noise or interference, requiring preprocessing to ensure accuracy and reliability. The specific preprocessing steps are as follows:

[0063] All parameters except for the rudder angle command are smoothed using a moving average filtering method to suppress high-frequency noise. The length of the moving average window is preset based on the physical characteristics and rate of change of each parameter. For example, the engine speed changes slowly, while the current and voltage change rapidly, so their window sizes are different.

[0064] All parameters are subjected to amplitude limiting to eliminate outliers caused by momentary sensor interference or malfunction. The amplitude limiting value is determined based on the historical maximum and minimum values ​​of each parameter, with a certain margin reserved to ensure the physical reasonableness of the data. This process effectively avoids the impact of extreme or unrealistic values ​​on data accuracy.

[0065] All collected data will be tagged with a uniform time stamp to ensure that data from the same time point comes from the same sampling period. Alignment operations ensure temporal consistency between data points and prevent errors caused by asynchronous data acquisition.

[0066] Through efficient data acquisition, processing, and preprocessing processes, reliable support can be provided for subsequent operating condition identification and optimization decisions, ultimately achieving efficient energy management across the entire system.

[0067] In one possible implementation, the temporal deep learning model used in this method is a classification model that combines a Long Short-Term Memory (LSTM) network with an attention mechanism. LSTM, as a special type of recurrent neural network (RNN), has strong time series data processing capabilities and can effectively capture long-term dependencies in the data. Combined with the attention mechanism, the model can not only focus on important time segments in the input sequence but also dynamically adjust its focus area according to the context, thereby processing multi-source heterogeneous time series data more accurately.

[0068] LSTM can learn the dynamic changes of time series data, and is especially suitable for capturing the complex time series characteristics of ships under different operating conditions, such as the changing trends of parameters such as main engine speed, ship speed, and wind speed over different time periods.

[0069] Attention mechanisms play a crucial role in sequence data processing, enabling models to autonomously decide which historical information to focus on at each time step. For the multi-condition adaptation problem in shipboard operations, parameter changes at certain moments can be critical to the final condition assessment. Attention mechanisms can help models automatically identify these key moments, improving their recognition capabilities and accuracy.

[0070] The model was pre-trained using time-series samples of multi-source heterogeneous operating parameters from historical datasets, along with corresponding manually labeled operating condition tags. The training data included ship operating parameters under different operating conditions, such as main engine speed, fuel injection quantity, and electrical load, and each time-series sample was manually labeled with the corresponding operating condition category.

[0071] By using the backpropagation algorithm, the cross-entropy loss between the model's predictions and the actual labels is minimized during training. Cross-entropy loss is a metric for measuring the difference between the model's predictions and the actual labels. Backpropagation optimizes the model parameters through gradient descent, allowing the model to gradually adjust its internal parameters to reduce prediction errors and thus better learn the relationship between the input data and the output labels.

[0072] The model, through backpropagation optimization, can effectively learn the nonlinear mapping relationship between complex multidimensional input sequences (including multiple different parameters such as engine speed, wind speed, and ship attitude) and refined operational condition classifications (such as normal operating conditions and overload conditions). This relationship reflects the connection between different ship operating states and specific operational condition categories, helping the system accurately determine the ship's current operating status.

[0073] After the model is trained, it is input with real-time multi-source heterogeneous operating parameter data, and the model will output a multi-dimensional vector. Each element in this multi-dimensional vector represents the confidence probability that the model determines the current state belongs to a certain predefined refined working condition. The confidence probability of each working condition category indicates the model's confidence level in that working condition; the higher the confidence level, the more likely the current state belongs to that working condition.

[0074] Finally, by comparing the confidence probabilities of each operating condition category in the multidimensional vector, the operating condition category with the highest confidence level is selected as the "dominant operating condition" for the current state. This dominant operating condition represents the model's best prediction of the ship's current operating condition and can provide a basis for subsequent energy management and optimization decisions.

[0075] By combining LSTM and attention mechanisms in a temporal deep learning model, the energy management and efficiency optimization of ship shaft generators can be significantly improved through efficient operating condition identification. The model's training and optimization process ensures efficient mapping from complex multidimensional input data to refined operating condition classifications, laying the foundation for intelligent ship management.

[0076] In one possible implementation, the refined operating condition categories defined in this method include several different navigation states. These operating conditions can be used to refine the ship's energy management and efficiency optimization strategies under different navigation environments. Specifically, they include the following operating conditions:

[0077] Rated speed, downwind and smooth seas: Under this condition, the main engine speed remains stable near its rated value, and the ship's speed is high with minimal fluctuations, indicating that the ship is operating under ideal downwind and downwind conditions. The ship's pitch and roll angles are low, indicating a smooth voyage, with wind and waves originating from the stern, and the electrical grid load is relatively stable. This condition is suitable for normal and efficient navigation.

[0078] Rated speed, headwind, waves, and heavy load navigation condition: This condition is characterized by the main engine speed remaining stable at its rated value, while the ship's speed is lower than expected. This typically indicates encountering significant headwinds and waves, affecting the ship's speed. Large pitch angles with a fixed period indicate that the ship requires greater stability in adverse weather conditions. At this time, the main engine's torque output is close to its maximum, and the fuel injection quantity is also at a high level, indicating that the ship is under a heavy load.

[0079] Deceleration and large rudder angle turning: Under these conditions, the main engine speed is decreasing or below its rated value, indicating that the ship is maneuvering or slowing down. A rudder angle command continuously deviating from zero indicates that the ship is making a large-angle turn, with continuous changes in speed and heading. Significant fluctuations in the power grid load may occur, indicating that under complex operating conditions, the system load changes more drastically.

[0080] The definition of refined operating conditions is not static, but rather derived through the analysis and processing of large amounts of historical data, combined with unsupervised clustering algorithms and domain expert knowledge. Unsupervised clustering algorithms are used to automatically extract different operating condition categories from ship operation data, particularly identifying complex patterns or similarities between groups hidden within the data. Expert knowledge then helps to further validate and refine these categories, ensuring that the definition of each operating condition category aligns with its performance in actual operation.

[0081] Unsupervised clustering algorithms can automatically identify patterns in data and cluster operating conditions with similar characteristics based on changes in ship operating parameters (such as rotational speed, speed, rudder angle, wind and waves). By classifying these operating condition categories, the operating status of ships can be effectively categorized, providing strong support for subsequent energy management and efficiency optimization.

[0082] Unsupervised clustering results can only provide the potential group structure in the data, while domain expert knowledge is used to understand these structures and further refine them. Experts can assess the rationality of each type of operating condition based on the actual operational experience of ships and revise them to ensure that the definition of refined operating condition categories meets actual navigation requirements.

[0083] By combining unsupervised clustering algorithms with expert knowledge, a set of refined ship operating conditions can be defined, providing a more intelligent and efficient framework for ship energy management and efficiency optimization, and significantly improving the ship's adaptability and energy efficiency in complex environments.

[0084] In one possible implementation, the following aspects need to be considered in order to calculate the overall efficiency:

[0085] Main unit fuel consumption rate: The efficiency of the main unit directly affects the overall fuel consumption. The fuel consumption rate of the main unit varies greatly under different load conditions, so the efficiency needs to be estimated based on the specific consumption of the main unit at the current load point.

[0086] Shaft-driven generator efficiency: The shaft-driven generator converts the ship's mechanical energy into electrical energy. Its efficiency is affected by factors such as load, operating speed and operating conditions, and needs to be considered in the model.

[0087] Auxiliary generator efficiency: The power generation efficiency of auxiliary generators under different load points on the ship should also be taken into account. The load conditions of different auxiliary machines will affect their power generation efficiency, thus affecting the supply and distribution of energy.

[0088] Energy storage system charging and discharging efficiency: The charging and discharging efficiency of energy storage systems is crucial for the utilization of ship energy, especially in situations of energy surplus or shortage. How to efficiently utilize energy storage systems can optimize the overall operation of the ship.

[0089] The specific form of the objective function will be adjusted according to the ship's current operating conditions. For example:

[0090] Excess power operation: When a ship has ample excess power, the optimization objective will shift towards maximizing energy capture. In this case, the system will focus on storing excess energy for use when power is scarce, or increasing energy collection by improving power generation efficiency.

[0091] Power Shortage Conditions: Under power shortage conditions, the system's optimization focus shifts to ensuring operational stability and main engine safety. In this situation, the system prioritizes the safe operation of the main engine to avoid overload or damage, while rationally allocating limited energy resources to ensure the ship's normal navigation.

[0092] The optimization process needs to dynamically adapt to changes in different operating conditions. During navigation, the ship's load and operating status will change, and the system needs to dynamically adjust the parameters and objective function of the optimization model through real-time monitoring and data feedback to cope with the constantly changing navigation environment and operating conditions.

[0093] In one possible implementation, the main engine is the core power source of the ship, and its reasonable load allocation directly affects the efficient use of energy. Under different operating conditions, the upper limit of the main engine load will be dynamically adjusted:

[0094] When a ship has ample spare power, the main engine's load limit is set higher to fully utilize available energy. This allows the ship to use as much of the main engine's power as possible without overloading.

[0095] Under power constraints, the main engine's load limit will be appropriately reduced to ensure that the main engine is not overloaded and to allow for more power margin in the propulsion system. This adjustment ensures that the ship can maintain safe and stable operation even under high load conditions.

[0096] Auxiliary generators play a crucial role in a ship's energy supply. Each auxiliary engine's efficiency characteristic curve exhibits a specific high-efficiency operating range; therefore, it is essential to ensure that the auxiliary engines operate within this high-efficiency range.

[0097] Based on the number of auxiliary generators currently in operation and the overall grid load, the system dynamically calculates the operating load of each auxiliary generator to ensure it operates within its most efficient range. This reduces energy loss and improves overall power generation efficiency.

[0098] Energy storage systems play a crucial role in regulating energy flow on ships. The state of charge (SBC) of these systems must be adjusted based on predictions of future operating conditions.

[0099] If it is predicted that ships will enter high-load operating conditions in the near future, the energy storage system will increase the target state of charge and store more energy to cope with the upcoming demand peak.

[0100] When it is predicted that a ship will be under low load for a period of time in the future, the target state of charge of the energy storage system will be appropriately lowered to ensure that it can absorb and store excess energy. This process helps to balance the ship's energy supply and demand and avoid energy waste.

[0101] Adaptive adjustment of constraint coefficients helps to optimize the operation of ship systems under varying operating conditions, improve overall energy efficiency, and effectively ensure the safety and stability of ship operation, ultimately reducing energy consumption and enhancing the ship's economy and sustainability.

[0102] In one possible implementation, firstly, based on the structure of the global efficiency optimization model, it is determined whether it is a convex optimization problem. Convex optimization problems have a unique optimal solution and usually possess favorable mathematical properties, allowing for rapid solution using precise mathematical methods. Specifically, when a ship's energy management model exhibits characteristics such as continuity and monotonicity, it can be identified as a convex optimization problem.

[0103] When the model is determined to be a convex optimization problem, the algorithm selection strategy will prioritize linear programming or quadratic programming algorithms. These two algorithms offer advantages in terms of fast computation speed and high solution accuracy, making them suitable for solving optimization problems with deterministic constraints. Linear programming is applicable when the system has a linear objective function and constraints, while quadratic programming can handle problems where the objective function is quadratic. These algorithms can quickly provide accurate optimal solutions, thereby improving the efficiency of ship energy management.

[0104] If the global efficiency optimization model contains non-convex components or discrete variables (such as the selection of certain operating states, switching control, etc.), then the algorithm selection strategy shifts to a genetic algorithm. A genetic algorithm is a heuristic search method that can effectively avoid getting trapped in local optima when solving complex nonlinear, non-convex optimization problems. By simulating the natural selection process, genetic algorithms adapt to environmental changes, quickly find suboptimal solutions to problems, and are suitable for handling large-scale or discrete problems.

[0105] The online optimization process monitors changes in the ship's operating conditions in real time and adjusts the optimization strategy based on real-time feedback. Regardless of the algorithm used, the optimization process is completed within a certain timeframe to meet the needs of engineering applications. Through online optimization, the ship can flexibly adjust its energy management strategy according to current operating conditions, maximizing system operating efficiency.

[0106] By dynamically selecting suitable optimization algorithms based on different operating conditions and model characteristics, the ship shaft generator energy management system can flexibly respond to changing operational needs, improve efficiency while ensuring system stability and safety, and provide strong support for efficient energy management of ships.

[0107] In one possible implementation, within each control cycle, multiple control commands are calculated based on the ship's operating conditions and energy requirements. These commands include:

[0108] The target active power command for a shaft-driven generator indicates the power that the shaft-driven generator should produce.

[0109] The instructions for starting and stopping auxiliary generators and the target power allocation determine the number of auxiliary generators that need to be started or stopped, as well as the power allocation of each unit.

[0110] The target charge / discharge power command for the energy storage system specifies whether the energy storage device should charge or discharge, and the amount of charge / discharge power.

[0111] Within a communication cycle, the aforementioned commands are sent to the corresponding execution controllers via the ship's control network (such as the ship's control bus or other real-time communication systems). Each controller is responsible for the execution of its corresponding energy subsystem; for example, the generator controller is responsible for regulating the generator's power, and the energy storage system controller is responsible for charging and discharging the battery.

[0112] Through coordination via the ship's control network, all calculated commands are sent within the same control cycle, requiring each controller to begin executing the new commands at the next unified control time starting point. In this way, all commands will begin execution at the same moment, ensuring that each subsystem starts or adjusts at the same time node, thereby guaranteeing the coordinated operation of the system.

[0113] Because control commands from multiple energy subsystems are executed simultaneously, coordinated action between these subsystems is ensured, preventing power oscillations or system instability caused by asynchronous commands. Each controller operates at a unified control time point after receiving a command, ensuring balanced power output from different energy sources and guaranteeing stable operation of the ship's systems.

[0114] The technology of synchronous control command issuance significantly improves the efficiency and stability of the ship's energy management system, providing a more reliable guarantee for the ship's energy use, while avoiding system instability problems that may be caused by asynchronous control commands.

[0115] In one possible implementation, the system continuously monitors and collects real-time data from multiple sources during ship operation, including:

[0116] Input multi-source heterogeneous operating parameters: for example, ship speed, load, navigation environment (sea state, weather, etc.).

[0117] The model outputs the following operating condition identification results: based on the analysis of the current operating conditions, the operating status of the ship is inferred.

[0118] The issued control commands and the resulting system state changes: information on the transmission of commands and changes in system state (such as engine output power, generator load, etc.). All this data is periodically stored in a buffer as new training samples.

[0119] At regular intervals, or when the number of new samples in the buffer reaches a preset size, the system will extract a portion of samples from the buffer for training. At this time, the model will perform incremental training based on this new data; that is, it updates and optimizes the model parameters based on the original model, rather than retraining from scratch. Specifically, this process includes:

[0120] Small learning rate: Training with a small learning rate prevents drastic adjustments to the model parameters. The purpose of this is to fine-tune the existing network weights so that the model can adapt to slow changes during ship operation, such as hull fouling and main engine performance degradation.

[0121] Adaptive adjustments: As training progresses, the model gradually adapts to subtle changes in ship operation, adjusting control strategies to ensure the continued effectiveness of energy management and efficiency optimization.

[0122] After each incremental training iteration, the updated model continues to function in ship operations, adjusting control commands in real time and monitoring actual execution results. New training data continues to be collected and stored, providing a foundation for the next round of incremental training, forming a closed loop of continuous optimization.

[0123] Periodic incremental training not only enhances the system's ability to cope with performance changes during long-term operation, but also ensures the efficient and stable operation of the ship under different working conditions through continuous adaptive optimization.

[0124] In one possible implementation, the system continuously collects operational data from the ship, including input parameters, output results, equipment status, and changes in operating conditions of the electrical system. This data, after preprocessing, is transmitted to the system's control unit. The data processing includes noise reduction, standardization, and real-time updates, providing an accurate basis for subsequent dynamic identification.

[0125] By analyzing real-time data, the system identifies the current operating conditions based on existing models. This process, through real-time matching and adjustment of input parameters, quickly determines the ship's current operating status. The results of dynamic operating condition identification form the basis for online optimization, helping the system predict future power demand and load changes.

[0126] Based on the results of operating condition identification, the system dynamically constructs an optimization model to formulate the optimal power generation strategy and energy allocation scheme. This model not only adjusts based on historical data and changes in operating conditions, but also predicts and prepares for possible future changes in operating conditions.

[0127] Based on the optimized model, the system optimizes energy management and control commands in real time. This process achieves real-time adjustment with the controller through a feedback mechanism, ensuring that each subsystem of the ship's electrical system (such as generators and energy storage devices) operates efficiently according to the current operating conditions, maintaining system stability and optimizing energy utilization.

[0128] The above four steps (data acquisition and processing, dynamic identification of operating conditions, dynamic construction of optimization models, and online optimization and control) will be executed periodically. This cycle time will be determined based on the dynamic response speed of the ship's electrical system. That is, when the system response is fast, the cycle time will be short to allow for more timely responses to changes in operating conditions; when the system response is slow, the cycle time can be appropriately extended to avoid excessively frequent control operations that could cause system instability or oscillations.

[0129] The periodic cyclic execution method of the closed-loop control system can ensure that the ship's power system responds to changes in operating conditions in a timely manner while avoiding instability caused by frequent control, thereby achieving efficient and stable energy management.

[0130] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0131] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for ship shaft power generation energy management and efficiency optimization based on multi-working condition adaptation, characterized in that, The method comprises the following steps: a data acquisition and processing step: real-time acquisition and preprocessing of multi-source heterogeneous operating parameters, including a main engine parameter group, a ship motion and environment parameter group, a power system parameter group, and a navigation instruction parameter group; a working condition dynamic identification step: inputting the processed main engine parameter group, ship motion and environment parameter group, power system parameter group, and navigation instruction parameter group into a pre-trained time series deep learning model to calculate a working condition identification result, which is used to represent the confidence degree of the ship being in a plurality of predefined fine working conditions at the current time, the working condition identification result including a dominant working condition, and the fine working condition being a typical operating state distinguished from a traditional two-way method according to multi-dimensional operating parameter combination clustering; an optimal model dynamic construction step: selecting and constructing a global efficiency optimization model according to the dominant working condition in the working condition identification result, the optimization objective of the global efficiency optimization model being to make the comprehensive efficiency of the whole path from the main engine fuel consumption to the power grid power output highest, the decision variable of the global efficiency optimization model including the power instruction of the shaft generator, auxiliary generator and energy storage system, and the constraint condition coefficient of the global efficiency optimization model being adaptively adjusted according to the dominant working condition; an online optimization and control step: solving the constructed global efficiency optimization model by using an online optimization algorithm, and synchronously issuing the optimal decision variable value obtained by solving to the execution mechanism of the shaft generator, auxiliary generator and energy storage system as a control instruction; an adaptive learning step: periodically incrementally training the time series deep learning model based on new data generated during system operation to update the network parameters thereof; the adaptive adjustment of the constraint condition coefficient refers to: the upper limit value of the main engine total load safe operation interval in the main engine load constraint, which is set to be higher in the working condition of abundant main engine surplus power to fully utilize energy, and lower in the working condition of tight main engine surplus power to reserve more power margin for the propulsion system; the high-efficiency operation interval range of a single auxiliary generator in the auxiliary generator operation constraint, which is dynamically calculated according to the number of currently running auxiliary generators and the total power grid load to ensure that the load rate of each running auxiliary generator falls within the high-efficiency section of its efficiency characteristic curve; the expected value of the target state of charge of the energy storage system in the energy storage system operation constraint, which is set according to the prediction of the working condition trend in the future period of time, and is adjusted to be higher to reserve energy if it is predicted to enter a high-load working condition, and is adjusted to be lower to prepare to absorb excess energy if it is predicted to enter a low-load working condition.

2. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method according to claim 1, characterized in that, The time series deep learning model is a classification model combining a long short-term memory network and an attention mechanism; The pre-training includes: training the model using a historical data set containing time series samples of the multi-source heterogeneous operating parameters and corresponding artificial labeled working condition labels, and minimizing the cross-entropy loss between the model prediction result and the true label in the training process by using a back propagation algorithm, so as to learn the nonlinear mapping relationship from the complex multi-dimensional input sequence to the fine working condition classification. The working condition recognition result is a multi-dimensional vector, and the value of each element in the vector represents the confidence probability that the model determines that the current state belongs to a certain predefined refined working condition. The dominant working condition is the working condition category with the maximum confidence probability value.

3. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method according to claim 2, characterized in that, The predefined refined working condition at least includes the following categories: Rated speed smooth sailing in following waves working condition: Its characteristics are that the main engine speed is stable near the rated value, the ship speed is high and has small fluctuations, the longitudinal and lateral angles have low amplitude, the wind and waves come from the stern direction, and the power grid load is stable; Rated speed heavy load sailing against the wind and waves working condition: Its characteristics are that the main engine speed is stable at the rated value, the ship speed is lower than expected, the longitudinal angle has a large amplitude and a fixed period, the wind and waves come from the bow direction, and the main engine output torque and fuel injection amount are close to the upper limit; Speed reduction maneuvering and large rudder angle turning working condition: Its characteristics are that the main engine speed is in the process of decreasing or lower than the rated value, the rudder angle command set value continuously deviates from zero, the ship speed and heading continuously change, and the power grid load can have large fluctuations; The category set of the refined working condition is inducted and defined by applying an unsupervised clustering algorithm to a large amount of historical data and combining domain expert knowledge.

4. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method of claim 1, wherein, In the dynamic construction step of the optimization model, the objective function of the global efficiency optimization model is a mathematical representation of the overall path comprehensive efficiency, which needs to consider the fuel consumption rate of the main engine at the current load point, the power generation efficiency of the shaft generator, the power generation efficiency of each auxiliary generator at its specific load point, and the charge and discharge efficiency of the energy storage system; The specific form of the objective function is adjusted according to the optimization focus implied by the dominant working condition.

5. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method of claim 1, wherein, In the online optimization and control step, the online optimization algorithm used is one of linear programming, quadratic programming, or genetic algorithm; the algorithm selection strategy is: when the global efficiency optimization model is judged to be a convex optimization problem, the linear programming or quadratic programming algorithm with fast calculation speed is preferentially selected for accurate solution; when the model contains non-convex or discrete variables, the genetic algorithm is selected for heuristic search to obtain a suboptimal solution that meets the engineering application requirements within an acceptable time.

6. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method of claim 1, wherein, The synchronous issuance of control instructions refers to sending the shaft generator target active power instruction, the auxiliary generator switching number and target power distribution instruction, and the energy storage system target charge and discharge power instruction calculated in a control period to the respective execution controllers through the ship control network in the same communication period, and requiring each controller to start executing the new instructions at the start of the next unified control time, so as to ensure the coordinated action of multiple energy subsystems and avoid power shock or system instability due to asynchronous instructions.

7. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method of claim 1, wherein, In the adaptive learning step, the specific process of periodic incremental training is as follows: the system continuously stores the actual data in the running buffer zone, including the input multi-source heterogeneous operating parameters, the model output working condition recognition results, the issued control instructions and the system state changes after execution; every fixed time period or when the number of new samples reaches a preset size, a part of the samples are extracted from the buffer zone to perform a round of incremental training on the time series deep learning model. During training, a small learning rate is used to fine-tune the network parameters, so that the model can slowly adapt to the drift of the running characteristics of the ship due to the fouling of the hull and the performance degradation of the main engine.

8. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method of claim 1, wherein, In the data acquisition and processing step, The acquired multi-source heterogeneous operating parameters specifically include: The main engine parameter group: the shaft speed and output torque of the main engine fuel engine are collected through the sensors installed on the main engine, and the fuel injection amount, scavenging pressure and exhaust temperature of the main engine are read through the main engine control system; The ship motion and environmental parameter group: the ship ground speed is obtained through the global satellite positioning system receiver, the pitch angle and roll angle are obtained through the attitude sensor installed near the center of gravity of the ship, the relative wind speed and direction are obtained through the wind speed and direction sensor installed at the top of the mast, and the wave height is estimated by analyzing the ship motion response and combining the model; The power system parameter group: the active and reactive power of the shaft generator is collected through the current and voltage transformers installed on the output circuit of the shaft generator, the total active load and bus voltage of the power grid are collected through the current and voltage transformers installed on the bus of the main distribution board, and the current state of charge of the energy storage system is read through the battery management system communication interface of the energy storage system; The navigation instruction parameter group: the current main engine speed instruction set value and rudder angle instruction set value are read from the bridge control system and rudder control system through the ship local area network; The preprocessing includes: first, using the moving average filtering method to smooth all parameters except the rudder angle instruction to suppress high-frequency noise, the time length of the moving average window is pre-set according to the physical characteristics and change speed of the parameters; second, using the amplitude limiting filtering method to process all parameters to eliminate abnormal values that are physically impossible due to sensor instantaneous interference, the threshold of the amplitude limiting is determined according to the historical statistical maximum and minimum values of each parameter with a certain margin; finally, all parameters are marked with a unified time scale and aligned to ensure that the data at the same time corresponds to the same sampling period.

9. The multi-operating condition adapted ship shaft power generation energy management and efficiency optimization method of claim 1, wherein, The method forms a closed-loop control system, and the data acquisition and processing step, the working condition dynamic identification step, the optimization model dynamic construction step and the online optimization and control step are periodically executed in a cycle, and the cycle period is determined according to the speed of the dynamic response of the ship power system.

Images