Method and device for controlling a towed fleet of large floating production platforms in a river-sea tow

Abstract

The present application relates to a kind of large-scale floating production platform river-sea towed tow formation control method and device, belong to marine engineering towed technical field, wherein, the large-scale floating production platform river-sea towed tow formation control method includes: real-time acquisition of the multi-source monitoring data of towed formation;Coupling feature extraction and feature fusion are carried out to the multi-source monitoring data of towed formation at current time to obtain target data;Target data is input to trajectory prediction model, and the output of trajectory prediction model is determined as the trajectory of towed formation at next time;Risk prediction is carried out based on the trajectory of towed formation at next time, and whether the pose of towed formation is controlled adjustment based on risk prediction result is determined.The present application carries out formation trajectory prediction and risk prediction by fusing deep learning method and distributed collaborative control theory, ensures the safe implementation of large-scale floating production platform from river to coastal water towed task.

Description

Technical Field

[0001] This invention relates to the field of marine engineering towing technology, and in particular to a control method and device for a large floating production platform towing formation in the river and sea. Background Technology

[0002] With the continuous advancement of offshore oil and gas resource development, large floating production platforms are being used more and more widely in the offshore oil and gas industry. These platforms are typically built and constructed along rivers and coastal ports, but because they lack their own power, they need to be towed by tugboats.

[0003] Rivers and narrow waterways have high vessel traffic and complex navigation environments, while coastal ports have large tidal ranges and dense fishing boat populations, making river-sea towing technology technically challenging, risky, and subject to numerous influencing factors. Existing floating production platform towing solutions are insufficient to meet the safety requirements of towing large floating production platforms from rivers to coastal waters. Summary of the Invention

[0004] In view of this, it is necessary to provide a control method and device for towing large floating production platforms in river and sea to solve the problem that existing floating production platform towing schemes cannot meet the safety requirements of towing large floating production platforms from rivers to coastal waters.

[0005] To address the aforementioned problems, in a first aspect, the present invention provides a control method for a large floating production platform towing convoy, comprising:

[0006] Real-time acquisition of multi-source monitoring data of the towing formation, including motion status parameters of the main tugboat and the large floating production platform, as well as navigation environment parameters;

[0007] The target data is obtained by coupling feature extraction and feature fusion of multi-source monitoring data of the towing formation at the current moment;

[0008] The target data is used as the input of the trajectory prediction model, and the output of the trajectory prediction model is determined as the trajectory of the towing formation at the next moment. The trajectory prediction model is based on the GRU-ADMM network to predict the trajectory of individual ships, and the trajectory prediction of the towing formation is performed through a distributed collaborative optimization algorithm. The trajectory prediction of individual ships includes the trajectory prediction of the main tug and the large floating production platform.

[0009] Risk prediction is performed based on the trajectory of the towing formation at the next moment, and the position and attitude of the towing formation are adjusted based on the risk prediction results.

[0010] In one possible implementation, the risk prediction based on the trajectory of the towed formation at the next moment includes:

[0011] By performing coupled feature extraction on the multi-source monitoring data of the towing formation at the next moment, the characteristics of towing behavior, environmental impact, ship motion state and traffic flow impact of the towing formation at the next moment are obtained.

[0012] Preprocessing is performed on the towing behavior characteristics, environmental impact characteristics, ship motion state characteristics, and traffic flow impact characteristics of the towing formation at the next moment to obtain the risk characteristic matrix of the towing formation at the next moment;

[0013] The risk feature matrix corresponding to the towing formation at the next moment is used as the input of the risk prediction model. Risk prediction is performed based on the risk value output by the risk prediction model, which is trained based on the historical risk feature matrix and historical risk value.

[0014] In one possible implementation, the risk value is calculated based on the following formula:

[0015]

[0016] in, Indicates the risk value. , , , As a weighting factor, This indicates the minimum ship spacing risk index. This represents the risk index for the shortest collision time. This indicates the sailing angle of the towing formation relative to other vessels. This indicates the tension of the towing cable.

[0017] In one possible implementation, determining whether to control and adjust the attitude of the towed formation based on risk prediction results includes:

[0018] If the risk value output by the risk prediction model is greater than the risk threshold, the attitude of the towing formation will be controlled and adjusted accordingly.

[0019] If the risk value output by the risk prediction model is less than or equal to the risk threshold, it will be determined not to control or adjust the attitude of the towing formation.

[0020] The control and adjustment of the towing formation's position and attitude includes: adjusting the tug configuration, track, and formation position and attitude of the towing formation.

[0021] In one possible implementation, the adjustment of the tugboat configuration of the towing formation includes:

[0022] Using the cost of towing missions as the objective function and thrust balance, torque balance, and equipment capacity limitations as constraint functions, the number of tugboats, tugboat power, and tugboat thrust direction of the towing formation are adjusted.

[0023] In one possible implementation, adjusting the number of tugboats, tugboat power, and tugboat thrust direction in the towing formation includes:

[0024] Using the number of tugboats, tugboat power, and tugboat thrust direction in the towing formation as variables, a globally optimal combination is generated based on a genetic algorithm.

[0025] The global optimal combination is locally optimized using a deep reinforcement learning network, and the locally optimized global optimal combination is further optimized using a genetic algorithm until the variable combination with the smallest objective function value is determined. The tugboat configuration of the towing formation is then adjusted based on the variable combination with the smallest objective function value.

[0026] In one possible implementation, the trajectory of the towed formation is adjusted, including:

[0027] Using total path cost as the objective function and channel width, speed, maneuverability, channel depth, and vessel encounter distance as constraint functions, the trajectory of the towing formation is adjusted based on the A* algorithm.

[0028] In one possible implementation, adjusting the formation attitude of the towed formation includes:

[0029] Construct a formation shape model for towing formations based on a chain leader-follower model;

[0030] The propulsion and steering torque of each vessel in the towing formation are adjusted based on the DDPG algorithm and the formation shape model corresponding to the towing formation.

[0031] In one possible implementation, the control and adjustment of the pose of the towed formation further includes:

[0032] The speed of each vessel is adjusted based on the real-time updated formation maneuver space of the towing formation;

[0033] The towline tension and towing force are calculated based on the advance and turning moment of each vessel in the real-time updated state space of the towing formation, and the towline length of each vessel is adjusted accordingly.

[0034] On the other hand, the present invention also provides a control device for a large floating production platform towing convoy, comprising:

[0035] The data acquisition module is used to collect multi-source monitoring data of the towing formation in real time. The multi-source monitoring data includes motion status parameters of the main tugboat and the large floating production platform, as well as navigation environment parameters.

[0036] The processing module is used to perform coupled feature extraction and feature fusion on the multi-source monitoring data of the towing formation at the current moment to obtain the target data;

[0037] The prediction module is used to take the target data as input to the trajectory prediction model and determine the trajectory of the towing formation at the next moment as the output of the trajectory prediction model. The trajectory prediction model is based on the GRU-ADMM network to predict the trajectory of individual ships and to tow formations through a distributed collaborative optimization algorithm. The trajectory prediction of individual ships includes the trajectory prediction of the main tug and the large floating production platform.

[0038] The control module is used to predict risks based on the trajectory of the towed formation at the next moment, and to determine whether to control and adjust the attitude of the towed formation based on the risk prediction results.

[0039] Secondly, the present invention also provides a control terminal, including a memory and a processor, wherein,

[0040] The memory is used to store programs;

[0041] The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the control method for the large floating production platform river-sea towing formation described in any of the above implementations.

[0042] Thirdly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the steps in the control method for the large floating production platform river-sea towing formation described in any of the above implementations.

[0043] The beneficial effects of this invention are as follows: The control method and device for towing a large floating production platform in river and sea, provided by this invention, obtains target data for trajectory prediction in the next moment by collecting multi-source monitoring data of the towing formation at the current moment, performing coupled feature extraction and feature fusion, and improving data accuracy to ensure the reliability of trajectory prediction. Then, based on the trajectory prediction model based on the GRU-ADMM network, a distributed cooperative optimization algorithm is used to predict the trajectory of the towing formation, providing accurate trajectory guidance for formation control and task execution. Finally, risk prediction is performed using multi-source data corresponding to the predicted trajectory, and the control adjustment of the towing formation is determined based on the risk prediction results to ensure the safety of the towing formation during towing operations. This invention, by integrating deep learning methods and distributed cooperative control theory for trajectory prediction and risk prediction, ensures the safe implementation of towing tasks of large floating production platforms from rivers to coastal waters. Attached Figure Description

[0044] Figure 1A schematic flowchart of an embodiment of the control method for a large floating production platform towing and towing formation provided by the present invention;

[0045] Figure 2 A schematic flowchart of an embodiment of the risk prediction and control process provided by the present invention;

[0046] Figure 3 A schematic flowchart of an embodiment of the tugboat configuration optimization process provided by the present invention;

[0047] Figure 4 A schematic flowchart illustrating an embodiment of the towing formation and operation process provided by the present invention;

[0048] Figure 5 A schematic diagram of an embodiment of the control device for a large floating production platform towing and hauling formation provided by the present invention;

[0049] Figure 6 A schematic diagram of an embodiment of the control terminal provided by the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0051] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0052] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0053] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0054] Before demonstrating the embodiments, the following terms will be explained.

[0055] Floating production platform: A floating production platform is a facility used for offshore oil and gas field development. It is capable of producing, processing and storing oil and gas at sea and is typically used in deep or offshore areas.

[0056] Floating production platform towing: Floating production platform towing refers to the process of towing a floating production platform from one location to another. This process typically involves complex operations and coordination to ensure the safety and stability of the platform.

[0057] This invention provides a control method and device for a large floating production platform towing convoy in the river and sea, which will be described below.

[0058] Figure 1 A schematic flowchart of an embodiment of the control method for a large floating production platform towing convoy provided by the present invention is shown below. Figure 1 As shown, the control method for large floating production platform river-sea towing formations includes:

[0059] S101. Real-time acquisition of multi-source monitoring data of the towing formation, including motion status parameters of the main tugboat and the large floating production platform, as well as navigation environment parameters.

[0060] It should be noted that the multi-source monitoring data of the towing formation can include motion status parameters and navigation environment parameters of the main tug and the large floating production platform. The motion status parameters can include data collected from the ship's Global Positioning System (GPS) sensors, Automatic Identification System (AIS), and Inertial Measurement Unit (IMU), while the navigation environment parameters can be data obtained from radar, weather stations, video surveillance stations, etc.

[0061] S102. Perform coupled feature extraction and feature fusion on the multi-source monitoring data of the towing formation at the current moment to obtain the target data.

[0062] It should be noted that the preprocessing of the multi-source monitoring data of the towing formation at the current moment may include operations such as outlier removal, timestamp unification, and coordinate system alignment to ensure data quality and consistency. Feature extraction from the preprocessed multi-source monitoring data can extract key features related to ship motion, mainly including: ship state features such as position, speed, heading angle, and rudder angle; formation coordination features such as towline tension, relative distances and angles between ships; and environmental features such as wind speed, current speed, wave height, and wave direction. During feature fusion, an attention mechanism can be used to weight the data from different sensors to dynamically allocate the importance of different data sources, and the time-series data and spatial topology can be fused using Kalman filtering. Kalman filtering can effectively reduce noise, improve data accuracy, and enable data from different sensors to work collaboratively.

[0063] S103. The target data is used as the input of the trajectory prediction model, and the output of the trajectory prediction model is determined as the trajectory of the towing formation at the next moment. The trajectory prediction model is based on the GRU-ADMM network to predict the trajectory of individual ships, and the trajectory prediction of the towing formation is performed through a distributed collaborative optimization algorithm. The trajectory prediction of individual ships includes the trajectory prediction of the main tug and the large floating production platform.

[0064] It should be noted that after obtaining the target data, it can be used as input to the trajectory prediction model, and the trajectory of the towing formation at the next moment can be determined by the output of the trajectory prediction model. The trajectory prediction model is based on a Gated Recurrent Unit (GRU) network and uses a distributed cooperative optimization algorithm for trajectory prediction. This invention introduces physical constraints based on the dynamic model, embedding physical factors such as ship-to-ship interactions and environmental influences into the training process of the GRU network. This method can effectively improve the model's predictive ability for complex dynamic systems and reduce the impact of missing data or noise on the results. The loss function of the trajectory prediction model can be expressed as: ,in , As a weighting factor, This represents the state prediction error based on the dynamic equations. This represents the mean square error between the model output and the actual trajectory. In predicting the motion of vessels in a formation, this invention further employs a distributed collaborative optimization algorithm to optimize the motion state of the entire formation. Each vessel performs local motion state prediction through edge computing nodes, and synchronously updates the global motion state based on the prediction results using a distributed optimization algorithm (such as the Alternating Direction Method of Multipliers, ADMM). This algorithm ensures that the motion trajectories of each vessel are coordinated, thereby improving the overall motion accuracy and consistency of the formation. Local prediction: Each vessel generates a short-term (e.g., 30-second) motion trajectory based on its own and the formation's state. Global optimization: The distributed algorithm coordinates the prediction results within the formation to ensure overall path consistency. Final motion trajectory output: Based on the results of global optimization, the final motion trajectory of the towing formation is output, providing precise trajectory guidance for formation control and mission execution.

[0065] S104. Based on the trajectory of the towed formation at the next moment, perform risk prediction, and determine whether to control and adjust the attitude of the towed formation based on the risk prediction results.

[0066] It should be noted that after determining the trajectory of the towing formation at the next moment, multi-source monitoring data of the towing formation at the next moment can be obtained for risk prediction. Then, the trajectory adjustment of the towing formation's position and attitude can be determined based on the risk prediction results, thereby ensuring the safety of the towing formation during the towing operation.

[0067] In summary, the control method for large floating production platforms towing formations in river and sea provided by this invention collects multi-source monitoring data of the towing formation at the current moment, performs coupled feature extraction and feature fusion to obtain target data for trajectory prediction at the next moment, and improves data accuracy to ensure the reliability of trajectory prediction. Then, based on the trajectory prediction model based on the GRU-ADMM network, a distributed cooperative optimization algorithm is used to predict the trajectory of the towing formation, providing accurate trajectory guidance for formation control and task execution. Finally, risk prediction is performed using multi-source data corresponding to the predicted trajectory, and the control adjustment of the towing formation is determined based on the risk prediction results to ensure the safety of the towing formation during towing operations. This invention integrates deep learning methods and distributed cooperative control theory for trajectory prediction and risk prediction, ensuring the safe implementation of large floating production platforms towing missions from rivers to coastal waters.

[0068] In some embodiments of the present invention, such as Figure 2 As shown, the specific process for risk prediction and control is as follows:

[0069] 1. Predicting the movement of towed vessel convoys based on distributed algorithms. The specific prediction scheme has been described above and will not be repeated here.

[0070] 2. Risk prediction is performed using a risk prediction model for towing fleets based on multi-source data fusion.

[0071] In some embodiments of the present invention, the risk prediction based on the trajectory of the towed formation at the next moment includes:

[0072] By performing coupled feature extraction on the multi-source monitoring data of the towing formation at the next moment, the characteristics of towing behavior, environmental impact, ship motion state and traffic flow impact of the towing formation at the next moment are obtained.

[0073] Preprocessing is performed on the towing behavior characteristics, environmental impact characteristics, ship motion state characteristics, and traffic flow impact characteristics of the towing formation at the next moment to obtain the risk characteristic matrix of the towing formation at the next moment;

[0074] The risk feature matrix corresponding to the towing formation at the next moment is used as the input of the risk prediction model. Risk prediction is performed based on the risk value output by the risk prediction model, which is trained based on the historical risk feature matrix and historical risk value.

[0075] Specifically, the risk prediction model for towing fleets based on multi-source data fusion mainly predicts the risks of grounding, collisions, and cable breaks by integrating AIS data, meteorological and hydrological data, and vessel status data. The specific prediction steps include:

[0076] (1) Data acquisition and preprocessing.

[0077] Data Acquisition: Collect multi-source data to be tested, including AIS data (speed, heading, position, etc.), meteorological and hydrological data (wind speed, current speed, tide, etc.), and ship status data (draft, towline tension, roll / pitch, etc.).

[0078] Data preprocessing: The multi-source data is preprocessed, including data denoising: Filtering algorithms (such as mean filtering, median filtering, wavelet denoising) or statistical methods are used to remove random noise from the data, improving data quality. Outlier detection: Interquartile range (INR) and Z-score (standard deviation method) are used to ensure data reliability. Standardization: Data of different dimensions are normalized (max-min normalization or Z-score normalization) to ensure scale consistency between features and improve model stability. Interpolation completion: For missing data, linear interpolation, spline interpolation, or Long Short-Term Memory (LSTM) networks based on temporal features are used to predict and complete missing values, ensuring data continuity and integrity. Fusion: Data from different sources (such as AIS data, meteorological data, tidal data, traffic flow data, etc.) are time-aligned, feature-matched, and fused into a unified data format for subsequent analysis and model training.

[0079] (2) Calculate the risk value based on the acquired dataset.

[0080] In some embodiments of the present invention, the risk value is calculated based on the following formula:

[0081]

[0082] in, Indicates the risk value. , , , As a weighting factor, This indicates the minimum ship spacing risk index. This represents the risk index for the shortest collision time. This indicates the sailing angle of the towing formation relative to other vessels. This indicates the tension of the towing cable.

[0083] (3) Obtain the input features and label values.

[0084] Based on the acquired risk values, an LSTM network is used to predict the risk of the towing fleet, obtaining input features and label values. Methods for obtaining input features include feature extraction, normalization, and data augmentation to ensure the stability and validity of the input data.

[0085] Input features can be categorized as follows:

[0086] Characteristics of towing:

[0087] Speed ​​(V): The speed of the tugboat convoy, which affects overall maneuverability and safety.

[0088] Towing cable tension (T): Fluctuations in towing cable tension reflect the stability of the towing process and are crucial to the coordinated movement of the vessel and safety risks.

[0089] Environmental impact characteristics:

[0090] Flow velocity (C): Changes in flow velocity affect the forces acting on and maneuverability of a tugboat fleet.

[0091] Tidal height (H): Tidal height determines the navigable depth and affects the navigation risk of tugboat fleets, especially in shallow waters or narrow channels.

[0092] Characteristics of ship motion state:

[0093] relative velocity ( The relative speed difference between the vessels in a tugboat convoy determines the coordination between the vessels and the risk of collision.

[0094] Relative direction ( The bearing angle of the towing vessel relative to the towed vessel or other vessels affects the vessel's collision avoidance strategy and risk level.

[0095] Traffic flow impact characteristics:

[0096] Ship density ( ): The number of ships per unit area or unit length of waterway, reflecting the impact of traffic flow on towing fleets.

[0097] Cross-navigation angle ( ( ): The angle between the tugboat convoy and the tracks of other vessels, typically used to assess collision risk. The smaller the angle, the higher the collision risk.

[0098] Ultimately, the input features can be represented as a risk feature matrix. .

[0099] (4) Construct a fleet risk prediction model based on XGBoost, and use the sliding window technique to continuously update the training samples in the time series data to dynamically capture the evolution trend of the towing fleet risk. Based on the trained XGBoost model, input the real-time multimodal information of the towing fleet, and calculate the potential risk of the towing fleet in the future time period in real time.

[0100] 3. Construct a risk prediction and risk management system for towing formations.

[0101] The system consists of a login module, a data receiving module, an electronic nautical chart module, a real-time vessel data module, a risk detection module, an early warning and emergency response module, a traffic control module, and a vessel scheduling and management module. Through the collaborative work of these multiple modules, it monitors navigation status in real time, predicts potential risks, and provides visual decision support.

[0102] In some embodiments of the present invention, determining whether to control and adjust the attitude of the towing formation based on risk prediction results includes:

[0103] If the risk value output by the risk prediction model is greater than the risk threshold, the attitude of the towing formation will be controlled and adjusted accordingly.

[0104] If the risk value output by the risk prediction model is less than or equal to the risk threshold, it will be determined not to control or adjust the attitude of the towing formation.

[0105] The control and adjustment of the towing formation's position and attitude includes: adjusting the tug configuration, track, and formation position and attitude of the towing formation.

[0106] Specifically, when the risk value output by the risk prediction model is greater than the risk threshold, it indicates that there is a safety risk in the towing formation continuing along its current trajectory, and therefore the attitude of the towing formation needs to be controlled and adjusted. When the risk value output by the risk prediction model is less than or equal to the risk threshold, it indicates that the towing formation can safely continue along its current trajectory, and no control adjustment is required.

[0107] In some embodiments of the present invention, adjusting the tugboat configuration of the towing formation includes:

[0108] Using the cost of towing missions as the objective function and thrust balance, torque balance, and equipment capacity limitations as constraint functions, the number of tugboats, tugboat power, and tugboat thrust direction of the towing formation are adjusted.

[0109] In some embodiments of the present invention, such as Figure 3 As shown, the tugboat configuration adjustment process of the towing formation is as follows:

[0110] I. Construction of Real-time Driving Force Calculation Model

[0111] The model involves the following key steps:

[0112] Dynamic model of platform and tugboat: Based on the platform's size, weight, draft, and the tugboat's power, traction force, and other parameters, an interaction model between the platform and the tugboat is established.

[0113] Impact of environmental factors: By acquiring environmental data in real time, such as wind speed, wind direction, and tidal current speed, the model will dynamically adjust the drag calculation results to ensure that it can cope with changes in the natural environment.

[0114] Channel and range factors: The width, depth, and curvature of the channel all affect the required drag force. Especially in narrow channels or bends, the calculation model will dynamically adjust the drag force according to the channel characteristics to prevent the platform from shifting laterally or running aground.

[0115] 1. Calculation of towing force and single-ship power throughout the process.

[0116] Total drag of towing Empirical calculation formula:

[0117]

[0118] In the formula This indicates the frictional resistance of the towed vessel. This indicates the remaining resistance of the tugboat. This indicates the air resistance of the towed vessel.

[0119] Approximate formula for calculating the resistance of a towed vessel:

[0120]

[0121]

[0122]

[0123] In the formula Indicates the underwater wetted area of ​​the towed vessel. This represents the cross-sectional area of ​​the portion of the vessel submerged in water. It indicates the towing speed (the combined speed of the ship and the water). Indicates wind speed. Represents the squareness coefficient. For air density, use 1.22 calculate, The shape factor represents the area exposed to wind.

[0124] 2. Dynamic real-time adjustment and optimization.

[0125] The model dynamically calculates towing demand by providing real-time feedback on wind force and towing speed data, and automatically adjusts tugboat power, traction distribution, and towing cable tension to ensure the stability and safety of towing operations.

[0126] In river-sea towing missions, when wind speed increases, the real-time calculation model can quickly assess the impact of increased wind force on platform drag and correspondingly increase the power output of the tugboats. On the other hand, if the tidal current velocity increases, the model will adjust the tugboat configuration according to the direction of the water flow to ensure the stability of the platform's navigation direction.

[0127] II. Cost Calculation Model for Towing Missions Based on Multi-Factor Optimization

[0128] Factors affecting the model:

[0129] Towing operations involve a variety of economic and technical factors that directly affect overall cost, efficiency, and safety. The following are the main influencing factors and their analysis:

[0130] Number of tugboats: The number of tugboats has a dual effect on towing costs: on the one hand, increasing the number of tugboats will directly increase the rental cost; on the other hand, reasonably increasing the number of tugboats can reduce the load on a single vessel, thereby reducing fuel consumption and equipment wear and tear, and achieving long-term cost savings.

[0131] Single-vessel power: Single-vessel power also has both positive and negative effects. High-powered tugboats are generally more expensive to charter, but they may save on overall costs by reducing the total number of tugboats required; however, high-powered equipment is often accompanied by higher fuel consumption, so a balance needs to be found between performance and economy.

[0132] Operating duration: The impact of operating duration on costs is mainly reflected in two aspects: first, rental fees are linearly positively correlated with time; second, fuel consumption increases with the accumulation of operating time. Therefore, rationally planning the operating cycle is key to reducing total costs.

[0133] Economic Factors: From an economic perspective, the impact of fuel price fluctuations and equipment efficiency on towing costs cannot be ignored. Rising fuel prices significantly increase operating costs, while equipment efficiency (such as the tugboat's energy efficiency ratio) directly affects energy consumption and cost control.

[0134] Technological constraints: From an economic perspective, the impact of fuel price fluctuations and equipment efficiency on towing costs cannot be ignored. Rising fuel prices significantly increase operating costs, while equipment efficiency (such as the tugboat's energy efficiency ratio) directly affects energy consumption and cost control.

[0135] In tugboat configuration optimization, the total cost typically consists of multiple components, including tugboat rental fees, fuel consumption costs, and operating time-related expenses. Therefore, it is necessary to obtain information on tugboat types, the number of potential tugboats, their power, fuel consumption, and rental fees.

[0136] (1) Tugboat rental costs.

[0137] Tugboat rental costs are directly related to the number of tugboats and the duration of operation, as shown in the following formula:

[0138]

[0139] In the formula, This represents the daily rental rate for the i-th type of tugboat. Indicates the duration of the task (in days). This indicates the number of tugboats.

[0140] (2) Fuel consumption cost.

[0141] Fuel consumption costs are related to tugboat power, operating time, and tugboat efficiency, as shown in the following formula:

[0142]

[0143] In the formula, This represents the fuel consumption coefficient per kilowatt of power. This represents the power of the i-th type of tugboat. Indicates fuel price, Indicates the duration of the task.

[0144] (3) Additional costs.

[0145] The additional costs take into account the impact of factors such as sea state changes, increased drag, and thrust redundancy on the total cost. Assuming a 10% increase in drag, fuel costs will rise by 8-12%; in complex sea conditions, thrust redundancy will increase by 15-20%. The formula for calculating the additional costs is as follows:

[0146]

[0147] In the formula, Indicates the percentage increase in resistance. This indicates the percentage increase in redundant thrust demand due to complex sea conditions.

[0148] (4) Cost of working hours.

[0149] Operating hours directly affect rental costs and fuel consumption, as shown in the following formula:

[0150]

[0151] In the formula, This indicates the fuel consumption rate.

[0152] Combining the above parts, we obtain the formula for calculating the total cost:

[0153]

[0154] The initial total cost is calculated using the above method. Subsequently, the tugboat configuration optimization model will select the optimal tugboat combination and adjust it according to real-time conditions. The total cost will also be updated based on real-time towing force, power, and fuel consumption.

[0155] III. Tug configuration optimization model considering trajectory control constraints of towing formations.

[0156] The core constraints of the model are divided into three parts: thrust balance equation, torque balance equation, and equipment capacity limit, which are designed to ensure the stability and safety of the towing formation.

[0157] (1) Thrust balance equation.

[0158] The thrust balance equation is one of the key constraints. Its function is to ensure that the total thrust provided by the tugboats can overcome the convoy's sailing resistance and maintain a certain safety margin of thrust to cope with complex environments (such as wind and current), thereby ensuring the convoy's smooth progress at the target speed. The conditions that the thrust balance equation must satisfy are:

[0159]

[0160]

[0161] In the formula, Indicates navigation resistance. Indicates safe redundancy thrust. Indicates the thrust direction angle of the tugboat. This indicates the current thrust of the ship.

[0162] (2) Torque balance equation.

[0163] The moment balance equations, by controlling the tugboat's arrangement coordinates and thrust direction angle, ensure a balanced distribution of thrust, thereby maintaining a stable heading and attitude for the towed platform. The conditions that the moment balance equations must satisfy are:

[0164]

[0165] In the formula, , This indicates the coordinates for the tugboat's arrangement.

[0166] (3) Equipment capacity limitations

[0167] Equipment capacity limits constrain the power of tugboats to prevent overloading of single or multiple tugboats, ensuring equipment safety and avoiding operational risks caused by overload. The conditions that equipment capacity limits must meet are:

[0168]

[0169] In the formula, This indicates the current maximum thrust of the ship.

[0170] In some embodiments of the present invention, adjusting the number of tugboats, tugboat power, and tugboat thrust direction of the towing formation includes:

[0171] Using the number of tugboats, tugboat power, and tugboat thrust direction in the towing formation as variables, a globally optimal combination is generated based on a genetic algorithm.

[0172] The global optimal combination is locally optimized using a deep reinforcement learning network, and the locally optimized global optimal combination is further optimized using a genetic algorithm until the variable combination with the smallest objective function value is determined. The tugboat configuration of the towing formation is then adjusted based on the variable combination with the smallest objective function value.

[0173] Specifically, the steps for adjusting the number of tugboats, tugboat power, and tugboat thrust direction in a towing formation include:

[0174] 1. First, obtain the relevant parameters for the towing mission, including the following parts:

[0175] Track control constraints: current position, target heading angle, maximum permissible yaw angle, maximum permissible track offset, channel curve radius, and current tugboat thrust direction.

[0176] Genetic Algorithm (GA) parameters: mutation rate, population size.

[0177] Deep Reinforcement Learning (DRL) parameters: learning rate, discount factor.

[0178] 2. Population initialization:

[0179] The initial population is generated by using the number, power, and thrust direction angle of each type of tugboat as variables:

[0180]

[0181] In the formula, This represents the number of tugboats of type i. This represents the power of the i-th type of tugboat. This represents the thrust direction angle of the i-th type of tugboat.

[0182] 3. Fitness function calculation:

[0183] The total cost, thrust error, and trajectory stability of each candidate scheme are evaluated, taking into account both economic efficiency and safety. The calculation formula is as follows:

[0184]

[0185] In the formula, Represents the total cost. This indicates the thrust error (the difference between the actual thrust and the required thrust). This indicates the stability of the flight path (torque balance error).

[0186] 4. Selection and optimization: Select suitable combinations of tugboat quantity, power and thrust direction through selection, cross-pollination and variation.

[0187] 5. Configure the Reinforcement Learning (DRL) Environment: In DRL training, the agent needs to interact with the environment and learn the optimal policy. Therefore, the reinforcement learning environment needs to be defined first. The reinforcement learning environment is configured based on the optimal combination selected by the genetic algorithm, mainly including:

[0188] (1) State space: The environmental state observed by the agent at each time t, including the current position, target heading angle, yaw angle, wind speed, water flow speed and tug thrust direction.

[0189] (2) Action space: The agent adjusts the direction of the tug thrust and the power of the tug.

[0190] (3) Reward function: DRL guides the agent to learn the optimal tugboat configuration strategy through the reward function, as shown in the following formula:

[0191]

[0192] In the formula, This indicates a fuel cost incentive. This indicates a reward for track stability. This indicates a thrust balance bonus.

[0193] 6. Model Training: The model training uses a deep reinforcement learning network to learn the relationship between the state and the optimal action.

[0194] 7. Gradient Descent Optimization: Through the gradient descent optimization process, a near-optimal tugboat configuration can be quickly found.

[0195] 8. Feedback Mechanism Optimization: After DRL optimization is completed, the optimization results are fed back to GA to update the fitness function, and the above steps are repeated.

[0196] 9. Execute optimization strategy: After sufficient training, the agent learns an optimal strategy and can select the optimal tug thrust distribution scheme according to different environmental conditions. In actual towing missions, it can adjust the tug thrust direction and optimize the towing power distribution at any time to adapt to changes in environmental factors such as wind speed and ocean currents.

[0197] In some embodiments of the present invention, such as Figure 4 As shown, the towing formation and operation process specifically includes adjustments to the route, formation, and attitude.

[0198] In some embodiments of the present invention, the trajectory of the towing formation is adjusted, including:

[0199] Using total path cost as the objective function and channel width, speed, maneuverability, channel depth, and vessel encounter distance as constraint functions, the trajectory of the towing formation is adjusted based on the A* algorithm.

[0200] Specifically, adjustments to the trajectory of the towing formation include:

[0201] 1. Based on inland waterway navigation rules, seaport navigation rules and navigation environment data, inland waterway network and seaport route network are constructed respectively. The maritime navigation environment is modeled and the maritime route network is generated by combining wind, wave and current information and using the A* algorithm.

[0202] 2. Based on complex network theory, a full-segment river-sea towing route network is constructed by integrating inland waterway network, seaport route network and maritime route network.

[0203] 3. Using an improved A* algorithm, the globally optimal navigation path is planned in the aforementioned river-sea towed route network. The improved A* algorithm includes the setting of the objective function and constraints, and the specific steps are as follows:

[0204] Objective function settings:

[0205]

[0206] In the formula This represents the total cost of the path. Let i be the length of the i-th waterway segment; The actual speed of the i-th segment of the waterway; The waiting time for segment i; This is a penalty term for the high-risk area (such as a rapid current) in the j-th segment; This is the risk weighting coefficient.

[0207] Constraint settings:

[0208] Channel width constraints:

[0209]

[0210] The usable width of the channel must be greater than the total width of the formation (including the safety margin Δ on both sides). The navigation area can be adjusted according to special rules. For example, large vessels with lower speeds within the designated navigation lanes should, as far as possible, navigate along the right edge of the navigation lane and may enter the recommended route if it is safe to do so.

[0211] Speed ​​constraints (taking the Jiangsu section of the Yangtze River as an example; different sections may have different constraints):

[0212] Maximum speed limit within the navigation lanes:

[0213]

[0214] Minimum speed limits within the navigation lanes:

[0215]

[0216] This restriction does not apply when navigating in other waters.

[0217] Manipulation constraints:

[0218]

[0219] Turning radius Three times the total length of the formation (to ensure that the tow cable is not subjected to excessive lateral tension).

[0220] Channel depth constraints:

[0221]

[0222] The dynamic water depth must meet the requirements of draft plus safety margin (δ=0.5m).

[0223] Rule compliance constraints:

[0224]

[0225] Taking inland waterway navigation as an example, the minimum meeting distance for all encountering vessels meets the standard. =0.3 nautical miles. When navigating along the coast, the minimum encounter distance can be adjusted appropriately based on the actual situation.

[0226] In some embodiments of the present invention, adjusting the formation attitude of the towing formation includes:

[0227] Construct a formation shape model for towing formations based on a chain leader-follower model;

[0228] The propulsion and steering torque of each vessel in the towing formation are adjusted based on the DDPG algorithm and the formation shape model corresponding to the towing formation.

[0229] Specifically, the formation posture adjustment process is as follows:

[0230] 1. Construct a tugboat dynamics model and a towing formation shape model, and initialize the formation shape matrix based on a chain leader-follower model. The initialization process of the formation shape model is as follows;

[0231]

[0232] in,( , ) represents the location. For speed, For heading angle, Angular velocity, and For propulsion and steering torque, For quality, For rotational inertia, and is the damping coefficient.

[0233] A chain-like leader-follower structure is adopted to define the formation matrix. :

[0234]

[0235] in, , This is the scaling factor. , The relative distance between the expectations of followers and direct leaders. For nodes His direct leader.

[0236] 2. Based on the planned navigation path and ship maneuvering rules, the state space and action space of the towed fleet are dynamically updated, and a formation-keeping reward mechanism is set up. The DDPG-LF method is used to achieve intelligent formation of the river-sea towed fleet. The specific steps of the DDPG-LF-based river-sea towing cooperative formation method are as follows:

[0237] At each time step t, each tugboat adjusts its current state accordingly. Select Action and receive rewards Store experience data in the classification replay pool, sample and update network parameters according to priority; if obstacles or turbulence are detected, trigger local path replanning and adjust reward weights.

[0238] Define the state space of the towed formation:

[0239]

[0240] Define the movement space of the towing formation:

[0241]

[0242] That is, changes in propulsion force and steering torque must satisfy physical constraints: .

[0243] Implementation of a dynamic formation algorithm for towed fleets based on DDPG:

[0244]

[0245] Formation maintenance rewards: , The current relative distance, Expected value; Energy consumption optimization: Path tracking reward: , This represents the maximum permissible track deviation.

[0246] In some embodiments of the present invention, the control and adjustment of the attitude of the towed formation further includes:

[0247] The speed of each vessel is adjusted based on the real-time updated formation maneuver space of the towing formation;

[0248] The towline tension and towing force are calculated based on the advance and turning moment of each vessel in the real-time updated state space of the towing formation, and the towline length of each vessel is adjusted accordingly.

[0249] Specifically, a multi-agent system is used to design the tugboat formation, where each tugboat acts as an independent agent, communicating with neighboring tugboats to maintain the target formation. The relative positions of the tugboats are adjusted using a consensus-based control law, as shown in the following formula:

[0250]

[0251]

[0252] In the formula, , It is about controlling the gain. , and , These are the positions of tugboats i and j. , and , These are the speeds of tugboats i and j.

[0253] Tension control is used to adjust the cable length. The optimal cable length is determined based on the tugboat's performance and environmental conditions: if the tension exceeds the limit, the cable length is shortened; if the tension is insufficient, the cable length is extended. The tension formula is:

[0254]

[0255] In the formula, It is the towing force applied by tugboat i. It is the cable length between the tugboat i and the floating platform.

[0256] To better implement the control method for the large floating production platform river-sea towing formation in the embodiments of the present invention, based on the control method for the large floating production platform river-sea towing formation, correspondingly, as follows: Figure 5 As shown, this embodiment of the invention also provides a control device for a large floating production platform river-sea towing formation. The control device 500 for the large floating production platform river-sea towing formation includes:

[0257] The acquisition module 501 is used to acquire multi-source monitoring data of the towing formation in real time. The multi-source monitoring data includes motion status parameters of the main tugboat and the large floating production platform, as well as navigation environment parameters.

[0258] Processing module 502 is used to perform coupled feature extraction and feature fusion on the multi-source monitoring data of the towing formation at the current moment to obtain target data;

[0259] The prediction module 503 is used to take the target data as the input of the trajectory prediction model and determine the output of the trajectory prediction model as the trajectory of the towing formation at the next moment. The trajectory prediction model is based on the GRU-ADMM network to predict the trajectory of individual ships and to predict the trajectory of the towing formation through a distributed collaborative optimization algorithm. The trajectory prediction of individual ships includes the trajectory prediction of the main tug and the large floating production platform.

[0260] The control module 504 is used to perform risk prediction based on the trajectory of the towed formation at the next moment, and to determine whether to control and adjust the attitude of the towed formation based on the risk prediction results.

[0261] The control device 500 for the large floating production platform river-sea towing formation provided in the above embodiments can realize the technical solutions described in the above embodiments of the control method for the large floating production platform river-sea towing formation. The specific implementation principles of each module or unit can be found in the corresponding content in the above embodiments of the control method for the large floating production platform river-sea towing formation, which will not be repeated here.

[0262] like Figure 6 As shown, the present invention also provides a control terminal 600. The control terminal 600 includes a processor 601, a memory 602, and a display 603. Figure 6 Only some of the components of the control terminal 600 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0263] In some embodiments, processor 601 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 602 or process data, such as the magnetic resonance image optimization method of the present invention.

[0264] In some embodiments, processor 601 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 601 may be local or remote. In some embodiments, processor 601 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, internal cloud, multi-cloud, etc., or any combination thereof.

[0265] In some embodiments, memory 602 may be an internal storage unit of control terminal 600, such as a hard disk or memory of control terminal 600. In other embodiments, memory 602 may also be an external storage device of control terminal 600, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on control terminal 600.

[0266] Furthermore, the memory 602 may include both internal storage units of the control terminal 600 and external storage devices. The memory 602 is used to store application software and various types of data installed on the control terminal 600.

[0267] In some embodiments, display 603 may be an LED display, a liquid crystal display, a touch-screen liquid crystal display, or an organic light-emitting diode (OLED) touchscreen. Display 603 is used to display information on control terminal 600 and to display a visual user interface. Components 601-603 of control terminal 600 communicate with each other via a system bus.

[0268] In one embodiment, when the processor 601 executes the control program for the large floating production platform river-sea towing convoy stored in the memory 602, the following steps can be implemented:

[0269] Real-time acquisition of multi-source monitoring data of the towing formation, including motion status parameters of the main tugboat and the large floating production platform, as well as navigation environment parameters;

[0270] The target data is obtained by coupling feature extraction and feature fusion of multi-source monitoring data of the towing formation at the current moment;

[0271] The target data is used as the input of the trajectory prediction model, and the output of the trajectory prediction model is determined as the trajectory of the towing formation at the next moment. The trajectory prediction model is based on the GRU-ADMM network to predict the trajectory of individual ships, and the trajectory prediction of the towing formation is performed through a distributed collaborative optimization algorithm. The trajectory prediction of individual ships includes the trajectory prediction of the main tug and the large floating production platform.

[0272] Risk prediction is performed based on the trajectory of the towing formation at the next moment, and the position and attitude of the towing formation are adjusted based on the risk prediction results.

[0273] It should be understood that when the processor 601 executes the control program for the large floating production platform river-sea towing formation in the memory 602, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0274] Furthermore, this embodiment of the invention does not specifically limit the type of the control terminal 600 mentioned. The control terminal 600 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic devices can also be other portable electronic devices, such as laptop computers with touch-sensitive surfaces (e.g., touch panels). It should also be understood that in some other embodiments of the invention, the control terminal 600 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0275] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions of the control methods for large floating production platforms towing and hauling formations provided in the above-described method embodiments.

[0276] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0277] The control method and device for large floating production platform towing formations in river and sea provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A control method for a large floating production platform river-sea towed tugboat convoy, characterized by, The method comprises the following steps: real-time acquisition of multi-source monitoring data of the towing formation, wherein the multi-source monitoring data comprises motion state parameters of the main tugboat and the large floating production platform and navigation environment parameters; coupling feature extraction and feature fusion of the multi-source monitoring data of the towing formation at the current time to obtain target data; input of the target data into a trajectory prediction model, determination of the output of the trajectory prediction model as the trajectory of the towing formation at the next time, and trajectory prediction of the towing formation by the distributed collaborative optimization algorithm based on the GRU-ADMM network for trajectory prediction of a single ship, wherein the trajectory prediction of the single ship comprises trajectory prediction of the main tugboat and the large floating production platform; risk prediction based on the trajectory of the towing formation at the next time, and determination of whether to control and adjust the pose of the towing formation based on the risk prediction result; the risk prediction based on the trajectory of the towing formation at the next time comprises: coupling feature extraction of the multi-source monitoring data of the towing formation at the next time to obtain towing behavior features, environmental influence features, ship motion state features and traffic flow influence features corresponding to the towing formation at the next time; pretreatment of the towing behavior features, the environmental influence features, the ship motion state features and the traffic flow influence features corresponding to the towing formation at the next time to obtain a risk feature matrix corresponding to the towing formation at the next time; input of the risk feature matrix corresponding to the towing formation at the next time into a risk prediction model, and risk prediction based on the risk value output by the risk prediction model, wherein the risk prediction model is trained based on historical risk feature matrices and historical risk values; the determination of whether to control and adjust the pose of the towing formation based on the risk prediction result comprises: in the case that the risk value output by the risk prediction model is greater than a risk threshold, it is determined that the pose of the towing formation is controlled and adjusted; in the case that the risk value output by the risk prediction model is less than or equal to the risk threshold, it is determined that the pose of the towing formation is not controlled and adjusted; the control and adjustment of the pose of the towing formation comprises adjustment of the number of tow wheels, the power of tow wheels and the direction of tow wheel thrust of the towing formation; the adjustment of the number of tow wheels, the power of tow wheels and the direction of tow wheel thrust of the towing formation comprises: adjustment of the number of tow wheels, the power of tow wheels and the direction of tow wheel thrust of the towing formation based on the genetic algorithm to generate a globally optimal combination; local optimization of the globally optimal combination based on the deep reinforcement learning network, and optimization of the globally optimal combination after the local optimization based on the genetic algorithm until a variable combination with the minimum objective function value is determined, and adjustment of the tow wheel configuration of the towing formation based on the variable combination with the minimum objective function value. the adjustment of the track of the towing formation comprises: ​ 2. The control method of a large floating production platform river-sea towing convoy according to claim 1, characterized in that, ​ The path total cost is taken as an objective function, and the channel width, speed, maneuverability, channel water depth, and ship encounter distance are taken as constraint functions.

3. The control method of a large floating production platform river-sea towing convoy according to claim 1, characterized in that, The adjusting of the formation pose of the towed formation includes: A formation shape model corresponding to the towed formation is constructed based on a chain leader-follower model; The propulsion force and steering torque of each ship in the towed formation are adjusted based on a DDPG algorithm and the formation shape model corresponding to the towed formation.

4. The control method of a large floating production platform river-sea towing convoy according to claim 1, characterized in that, The control adjustment of the pose of the towed formation further includes: The speed of each ship is adjusted based on the real-time updated formation action space of the towed formation; The towline length of each ship is adjusted based on the calculation of the towline tension and tow force of each ship in the real-time updated state space of the towed formation.

5. A control device for a large floating production platform river-sea tug towed convoy, characterized by, It includes: The acquisition module is configured to acquire multi-source monitoring data of the towed formation in real time, wherein the multi-source monitoring data includes motion state parameters of the main tugboat and the large floating production platform and navigation environment parameters; The processing module is configured to perform coupling feature extraction and feature fusion on the multi-source monitoring data of the towed formation at the current time to obtain target data; The prediction module is configured to input the target data into a trajectory prediction model and determine an output of the trajectory prediction model as a trajectory of the towed formation at the next time, wherein the trajectory prediction model is based on a GRU-ADMM network to perform trajectory prediction of a single ship and perform trajectory prediction of the towed formation through a distributed collaborative optimization algorithm, and the trajectory prediction of the single ship includes trajectory prediction of the main tugboat and the large floating production platform; The control module is configured to perform risk prediction based on the trajectory of the towed formation at the next time and determine whether to perform control adjustment of the pose of the towed formation based on a risk prediction result. The risk prediction based on the trajectory of the towed formation at the next time includes: Coupling feature extraction is performed on the multi-source monitoring data of the towed formation at the next time to obtain towage behavior features, environmental influence features, ship motion state features, and traffic flow influence features corresponding to the towed formation at the next time; The towage behavior features, the environmental influence features, the ship motion state features, and the traffic flow influence features corresponding to the towed formation at the next time are preprocessed to obtain a risk feature matrix corresponding to the towed formation at the next time; The risk feature matrix corresponding to the towed formation at the next time is input into a risk prediction model, and risk prediction is performed based on a risk value output by the risk prediction model, wherein the risk prediction model is trained based on historical risk feature matrices and historical risk values; The determination of whether to perform control adjustment of the pose of the towed formation based on the risk prediction result includes: In a case where the risk value output by the risk prediction model is greater than a risk threshold, it is determined to perform control adjustment of the pose of the towed formation; In a case where the risk value output by the risk prediction model is less than or equal to the risk threshold, it is determined not to perform control adjustment of the pose of the towed formation; The control adjustment of the pose of the towed formation includes adjustment of tow propeller configuration, a path, and a formation pose of the towed formation. The adjustment of the tow propeller configuration of the towed formation includes: Taking a towing task cost as a target function, taking thrust balance, torque balance and equipment capacity limit as constraint functions, the number of tugboats, the power of tugboats and the thrust direction of tugboats of a towing formation are adjusted; The adjusting the number of tugboats, the power of tugboats and the thrust direction of tugboats of a towing formation comprises: Taking the number of tugboats, the power of tugboats and the thrust direction of tugboats of a towing formation as variables, a global optimal combination is generated based on a genetic algorithm; The global optimal combination is locally optimized based on a deep reinforcement learning network, and the global optimal combination after local optimization is optimized based on a genetic algorithm until a variable combination with a minimum target function value is determined, and the configuration of tugboats of a towing formation is adjusted based on the corresponding variable combination with the minimum target function value.

Images