Energy management optimization method for new energy intelligent yacht power system

Abstract

The invention relates to an energy management optimization method for a new energy intelligent yacht power system, and belongs to the technical field of energy management optimization. The method comprises the following steps: acquiring a tracking path: generating the tracking path by selecting a target point according to a current position and a target path of a yacht; acquiring a tracking rudder angle: acquiring a reference point selection distance according to a control fineness requirement and an energy state, selecting a group of rudder angle control reference points from a tracking path starting point, and calculating to obtain a steering engine rotation angle; obtaining a power demand: calculating and obtaining the power demand of the yacht according to the steering engine rotation angle and the navigational speed of the yacht; and yacht energy management: obtaining an energy management strategy based on the power demand under the yacht operation condition through deep reinforcement learning. According to the invention, through closed-loop linkage of path planning, rudder angle control, power calculation and energy distribution, bidirectional cooperation of a navigation task and energy response is realized.

Description

Technical Field

[0001] This invention belongs to the field of energy management optimization technology, and specifically relates to an energy management optimization method for a new energy intelligent yacht power system. Background Technology

[0002] As a recreational mode of transportation, yachts require high levels of energy efficiency, range, and route tracking accuracy in their power systems. Energy management of hybrid power systems is a key factor in improving the stability of a yacht's energy supply and extending its range.

[0003] However, existing yacht propulsion systems and energy management cannot integrate path tracking with energy management, and fail to dynamically adjust energy allocation in real time according to the sailing status. Energy management methods mostly rely on preset fixed rules or empirical parameters, which cannot cope with complex and dynamically changing sailing tasks, especially since variables such as power demand, solar energy, and battery charge and discharge status during path tracking are not fully considered. Summary of the Invention

[0004] To address the aforementioned problems in the existing technology, this invention provides an energy management optimization method for a new energy intelligent yacht power system. By linking path planning, rudder angle control, power calculation, and energy distribution in a closed loop, a two-way synergy between navigation tasks and energy response is achieved.

[0005] The objective of this invention can be achieved through the following technical solutions: The first aspect of this disclosure provides an energy management optimization method for a new energy intelligent yacht power system, comprising the following steps: Obtain the tracking path: Generate a tracking path by selecting target points based on the yacht's current position and the target path; Obtain the tracking rudder angle: Based on the control fineness requirements and energy status, select the spacing of the reference points, select a set of rudder angle control reference points starting from the starting point of the tracking path, and calculate the servo angle; Obtain power requirements: Calculate the yacht's power requirements based on the yacht's steering gear angle and speed; Yacht Energy Management: Energy management strategies are derived from power requirements under yacht operating conditions using deep reinforcement learning; The calculated servo rotation angle includes: Based on the yacht's current energy state, the reference point selection distance Δ is calculated. s : ; In the formula, S 0 is the reference point spacing. c The energy influence coefficient. E S Energy tension coefficient; Based on the deviation between the yacht's current position and the reference point, it is converted into the deflection angle of the rudder. Then, based on the current operating data, the deflection angle is corrected for errors to obtain the final tracking rudder angle.

[0006] Furthermore, the error correction of the deflection angle includes: A dataset is built using historical navigation data. After preprocessing the collected raw data, features for prediction are extracted from the raw data using the correlation coefficient method. Constructing a BP neural network model: Set the number of nodes in the input layer according to the selected prediction features, set the output layer to compensate for bias, and allocate the collected data into training set, validation set and test set to complete the model training; Based on the predictive features selected through model training, corresponding sensors are set up to collect operating condition data in real time. The collected data is then input into the machine learning model to predict and output the rudder angle to compensate for the deviation, thereby correcting the error in the deflection angle.

[0007] Furthermore, obtaining the tracking path includes the following steps: Selecting target points: Sample a set of candidate target points from the target path, filter the candidate target points by traversing the candidate target points to select the target points that meet the threshold range of direction difference and curvature difference, and take the point closest to the current position as the final target point; Generate initial tracking path: Based on the ship's current position and the selected target point, the initial tracking path of the yacht is randomly generated using B-spline control points; Searching for control points: The control points of the B-spline curve are encoded, and the optimal control point configuration is searched through the evolutionary process of a genetic algorithm to achieve the optimal tracking path planning goal.

[0008] Furthermore, the generation of the initial tracking path also includes the step of: Let the time domain be The number of control points is n The B-spline curve locus is: ; In the formula, For the spline curve locus, t ∈ ; These are the control points for the B-spline curve. For the first i indivual k B-order spline basis functions, basis functions It is constructed using the de Boer-Cox recurrence relation, and its expression is: ; In the formula, U 1. U2 is a basis function The coefficient; Introducing B-spline curves into the representation of ship paths, a set of control point coordinates can be expressed as: ; The path expression generated by this set of control points is: ; In the formula, X ( t ) indicates that a point on the path is at x Changes in position on the axis. Y ( t ) indicates that a point on the path is at y The positional changes on the axis.

[0009] Furthermore, the search for control points also includes the step of: Encode the control points and solve for the control points ( , )coordinate: ; In the formula, ( , ( ) is the starting point of the path s and the end point f Connection sf The coordinates of the division points, Starting point s and the end point f The distance between them rand are random numbers and rand ∈[-0.5,0.5], It is a straight line sf The slope; Establish the fitness function: Based on the actual needs in the error correction process, set constraints to establish the fitness function. The formula for the fitness function is: ; In the formula, , , These are the weighting coefficients. The path distance. For path smoothness, The path direction should match the objective function; Genetic Iterative Optimization: Minimize the fitness function through iterative optimization to obtain the optimal path.

[0010] Furthermore, the energy tension coefficient E S The calculation formula is: ; In the formula, SOC The battery is in its state of charge. P pv The average power of photovoltaic power. k for SOC Sensitivity coefficient α This is the photovoltaic weighting coefficient.

[0011] Furthermore, the acquisition of power requirements includes the following steps: Propulsion power is obtained based on the ship's motion. and dynamic compensation power Δ P d The calculation formula is: ; In the formula, r For the density of water, C d This is the hull drag coefficient. A This represents the projected area of ​​the ship's hull in contact with the water. v c At the current speed, C w The coefficient of water resistance. i The pitch angle; ; In the formula, a For acceleration; Yacht steering power is obtained from the yacht's rudder angle. The calculation formula is: ; In the formula, The torque coefficient, d For servo motor turning angle, oh The angular velocity of the rudder blade. or For the efficiency of the hydraulic system, k l This is the steering loss coefficient; Based on propulsion power, dynamic compensation power, and yacht steering power, calculate the yacht's total power requirement: P z = +Δ P d + ; In the formula, P z This represents the total power requirements of the yacht.

[0012] Furthermore, the yacht energy management includes the following steps: Define the state space: Construct the state space S of the hybrid yacht by including the yacht's state detection variables and all state variables required for path tracing.t : S t =[PV t SOC t P zt , v ct ]; In the formula, PV t Let be the photovoltaic power at time t, and SOC. t Let P be the battery state of charge at time t. zt Let t be the power demand. v ct The speed is the commanded speed at time t; Setting the Action Space: Based on the generator and power battery, setting the action space A of the hybrid yacht. t : A t =[DE t BP t ]; In the formula, DE t For the generator output, BP t The power used to charge or discharge the power battery; The constraints include: ; ≤DE t ≤min( DE max , + ); In the formula, BP max Δ is the maximum allowable discharge power of the battery. P d For dynamic power compensation, DE max This is the maximum allowable output power of the generator. For yacht propulsion power, For yacht steering power; Set the reward function: Construct a reward function R based on the power requirements of the target path tracking and the target speed. R=- l 1(P SOC +P BP +P z )+ l 2( ); In the formula, l 1 represents the penalty weighting coefficient. l 2 represents the reward weighting coefficient, P SOC For the state of charge penalty, P BP For charge / discharge power penalty, Pz For path tracing, power penalty is required, v r For actual speed, v c This is the command speed.

[0013] Furthermore, the yacht energy management also includes the following steps: Network training: A deep Q-network is used for training. After initializing the network, iterative training is performed, and training termination conditions are set to complete the network training. Energy management strategy acquisition: State variables are acquired through sensors and path tracking data, and after normalization, they are input into the network to acquire actions. Constraint checks are performed on the actions and out-of-bounds actions are corrected. The generator and battery are controlled to output power according to the corrected actions.

[0014] The second aspect of this disclosure provides a new energy intelligent yacht power system, which implements the energy management optimization method of the new energy intelligent yacht power system as described above, including a photovoltaic cell pack, a generator set, a power battery, an electric motor, and a load; The photovoltaic cell array is used to convert solar energy into electrical energy, which is connected to a DC bus via a DC / DC converter to directly power the load or charge the battery. The generator set is used to provide stable, high-power power output, and the output AC power is rectified by an AC / DC converter and then connected to a DC bus. The power battery is used for energy storage and buffering, balancing supply and demand fluctuations, providing short-term high power output, and is connected to a DC bus via a bidirectional DC / DC converter to support charging and discharging. The electric motor and load are used to convert electrical energy into mechanical energy to drive the yacht's propulsion system.

[0015] The beneficial effects of this invention are as follows: This invention first introduces the current energy state of the yacht as a reference point selection interval during path tracing, thereby optimizing the parameters of the rudder angle for path tracing to accommodate the energy state considerations in the path tracing process of energy management in this embodiment. Then, the power demand during path tracing is incorporated into the energy management of the hybrid yacht based on DRL, enabling the energy management system to "anticipate" future sailing needs and make more reasonable energy allocation decisions. By achieving closed-loop linkage of path planning, rudder angle control, power calculation, and energy allocation, a two-way synergy between "sailing mission and energy response" is realized, solving the contradiction between energy consumption and accuracy of new energy yachts in complex environments. This achieves dynamic synergy between path tracing and energy management, providing a high-efficiency power control paradigm for intelligent yachts. Attached Figure Description

[0016] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0017] Figure 1 This is a schematic diagram illustrating the steps of an energy management optimization method for a new energy intelligent yacht power system provided in an embodiment of the present invention. Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0019] This embodiment provides an energy management optimization method for a new energy intelligent yacht power system, such as... Figure 1 As shown, it includes the following steps: S1. Obtain the tracking path: Generate a tracking path by selecting target points based on the yacht's current position and the target path, including the following steps: S11. Select target point: Sample a set of candidate target points from the target path, filter the target points that meet the threshold range of direction difference and curvature difference by traversing the candidate target points, and take the point closest to the current position as the final target point; It should be noted that when generating the tracking path, a certain point on the target path needs to be determined as the endpoint of the tracking path. However, the selection of this target point requires that the path from the ship's current position to the target point achieves the best match with the target path in terms of curvature and direction.

[0020] S12. Generate initial tracking path: Based on the ship's current position and the selected target point, randomly generate the yacht's initial tracking path using B-spline control points, including the following steps: Let the time domain be The number of control points is n The B-spline curve locus is: ; In the formula, For the spline curve locus, t ∈ ; These are the control points for the B-spline curve. For the first i indivual k B-order spline basis functions, basis functions It is constructed using the de Boer-Cox recurrence relation, and its expression is: ; In the formula, U 1. U 2 is a basis function The coefficient.

[0021] Introducing B-spline curves into the representation of ship paths, a set of control point coordinates can be expressed as: ; The path expression generated by this set of control points is: ; In the formula, X ( t ) indicates that a point on the path is at x Changes in position on the axis. Y ( t ) indicates that a point on the path is at y The positional changes on the axis.

[0022] It should be noted that in this embodiment, the shape and direction of the spline curve are adjusted using control points of the B-spline curve, and the first and last control points are the start and end points of the spline curve trajectory, respectively. , B-spline interpolation optimization method can interpolate path points at certain intervals, resulting in smoother curve paths, more continuous curvature changes, and improved path quality and characteristics.

[0023] S13. Searching for Control Points: The control points of the B-spline curve are encoded, and the optimal control point configuration is searched through the evolutionary process of a genetic algorithm to achieve the optimal tracking path planning objective. This includes the following steps: Encode the control points and solve for the control points ( , )coordinate: ; In the formula, ( , ( ) is the starting point of the path s and the end point f Connection sf The coordinates of the division points, Starting point s and the end point f The distance between them rand are random numbers and rand ∈[-0.5,0.5], It is a straight line sf The slope.

[0024] Understandably, the starting point s and the end point f The distance between them can be determined based on the starting point. s and the end point fThe latitude and longitude coordinates are solved using the distance formula. Through the above representation, the control points of the B-spline curve are represented as individuals of the genetic algorithm. That is, a set of control points represents a feasible path, and a tracking path represents a genetic individual, thus completing the subsequent optimization.

[0025] Establish the fitness function: Based on the actual needs in the error correction process, set constraints to establish the fitness function. The formula for the fitness function is: ; In the formula, , , These are the weighting coefficients. The path distance. For path smoothness, The objective function is used to fit the path direction. Where: ; In the formula, Representing a path z Upper i discrete points, h The number of discrete points is used to constrain the path distance, minimizing the total distance of the tracing path.

[0026] ; In the formula, Indicates the first path i The direction angle of the nth point, it passes through the nth point i Point and the i Calculated by the relative position between +1 points. This represents the curvature change between three adjacent path points. To make the path as smooth as possible, the curvature should be as small as possible.

[0027] ; In the formula, For the ship on the tracking path i The direction and angle of the point For the first on the target path i The direction and angle of the point.

[0028] By minimizing This allows the actual planned tracking path to closely match the direction of the target path, avoiding deviation from the expected direction of the target path.

[0029] Genetic Iterative Optimization: Minimize the fitness function through iterative optimization to obtain the optimal path.

[0030] S2. Obtain the tracking rudder angle: Based on the control granularity requirements and energy status, select the spacing of the reference points, choose a set of rudder angle control reference points starting from the beginning of the tracking path, and calculate the servo angle; including the following steps: Based on the yacht's current energy state, the reference point selection distance Δ is calculated. s : ; In the formula, S 0 is the reference point spacing. c The energy influence coefficient. E S This represents the energy tension coefficient.

[0031] Understandably, when energy is scarce, the current energy state of the yacht increases the density of reference points to improve control accuracy and avoid additional energy consumption due to path deviation; when energy is sufficient, the number of reference points is reduced to lower the computational load and save processing power. This process optimizes the parameters of the path tracking rudder angle by influencing the selected reference points, so as to adapt to the energy state considerations of the path tracking process in energy management in this embodiment, and realize the two-way coordination of path tracking and energy management.

[0032] Among them, the energy tension coefficient E S The calculation formula is: ; In the formula, SOC The battery is in its state of charge. P pv The average power of photovoltaic power. k for SOC Sensitivity coefficient α This is the photovoltaic weighting coefficient.

[0033] Understandably, the formula for calculating the energy stress coefficient uses the Sigmoid function to achieve a nonlinear response by setting an energy stress state when the SOC is below 30%.

[0034] Based on the deviation between the yacht's current position and the reference point, it is converted into the deflection angle of the rudder. Then, based on the current operating data, the deflection angle is corrected for errors to obtain the final tracking rudder angle.

[0035] The error correction for the deflection angle includes: A dataset is built using historical navigation data, including but not limited to current position, target position, rudder angle, speed, heading, wind speed, wind direction, and ocean current information. After preprocessing the collected raw data, features for prediction are extracted from the raw data using the correlation coefficient method. Constructing a BP neural network model: Set the number of nodes in the input layer according to the selected prediction features, set the output layer to compensate for bias, and allocate the collected data into training set, validation set and test set in a ratio of 70%, 15% and 15% to complete the model training. Based on the predictive features selected through model training, corresponding sensors are set up to collect operating condition data in real time. The collected data is then input into the machine learning model to predict and output the rudder angle to compensate for the deviation, thereby correcting the error in the deflection angle.

[0036] S3. Obtain Power Requirements: Calculate the yacht's power requirements based on the yacht's steering gear angle and speed, including the following steps: S31. Obtain propulsion power based on the ship's motion state. and dynamic compensation power Δ P d The calculation formula is: ; In the formula, r For the density of water, C d This is the hull drag coefficient. A This represents the projected area of ​​the ship's hull in contact with the water. v c At the current speed, C w The coefficient of water resistance. m This represents the total mass of the yacht, where g is the acceleration due to gravity. i The heel angle is the angle of slope caused by the difference in draft between the bow and stern of the ship.

[0037] ; In the formula, a It is acceleration.

[0038] It should be noted that propulsion power includes water resistance power and gradient / load power. Dynamic compensation power describes the additional power compensation required by the ship due to acceleration / deceleration, which is used to overcome the influence of inertial forces on the propulsion system.

[0039] S32. Obtain yacht steering power based on yacht rudder angle. The calculation formula is: ; In the formula, The torque coefficient, d For servo motor turning angle, oh The angular velocity of the rudder blade. or For the efficiency of the hydraulic system, k l This is the steering loss coefficient.

[0040] It should be noted that the steering of a cruise ship is usually achieved through a steering system or a rotatable propeller. In the calculation of the steering power of a cruise ship, hydraulic power is the core energy source driving the steering system.

[0041] S33. Based on propulsion power, dynamic compensation power, and yacht steering power, calculate the yacht's total power demand P. z : P z = +Δ P d + ; It should be noted that the reason why the power error caused by environmental interference was not considered when calculating the total power demand of the yacht in this embodiment is that the environmental interference error has been included in the rudder angle during the process of obtaining the rudder angle, and is included in the calculation of steering power. At the same time, the environmental interference error is included in the correction of the rudder angle in advance to ensure that the error is eliminated in the path tracking process.

[0042] S4. Yacht Energy Management: Energy management strategies are derived based on the yacht's power requirements under operating conditions using deep reinforcement learning. This includes the following steps: S41. Define the state space: Construct the state space of the hybrid yacht by combining the yacht's state detection variables and all state variables required for path tracing. S t =[PV t SOC t P zt , v ct ]; In the formula, PV t Let be the photovoltaic power at time t, and SOC. t Let P be the battery state of charge at time t. zt Let t be the power demand. v ct Let t be the commanded speed.

[0043] It should be noted that in this embodiment, the power demand and commanded speed from path tracing are added to the state space, thereby introducing power demand and future demand (commanded speed) to expand the state space, enabling energy management to "anticipate" the needs of the navigation mission and thus make more reasonable energy allocation decisions.

[0044] S42. Setting the motion space: Based on the generator and power battery, setting the motion space of the hybrid yacht: A t =[DE t BP t ]; In the formula, DE tFor the generator output, BP t The power used to charge or discharge the battery.

[0045] The constraints include: ; ≤DE t ≤min( DE max , + ); In the formula, BP max This refers to the battery's maximum permissible discharge power. DE max This represents the maximum permissible output power of the generator.

[0046] It should be noted that the lower limit for the power of charging or discharging the power battery is if the photovoltaic power PV t Insufficient to cover dynamic needs ( The battery needs to make up the difference, if the photovoltaic power is sufficient ( The battery can remain undischarged (BP). t ≥0), the upper limit of the battery discharge power must not exceed its maximum allowable value. During this process, renewable energy sources such as photovoltaics are given priority, and batteries are used as a supplement. Regarding the output constraints of the generator, the lower limit of the generator must at least meet the steering power requirements, and the upper limit of the generator must not exceed its maximum rated power and must not exceed the current total propulsion + steering power requirements (to avoid overload waste).

[0047] S43. Set the reward function: Construct a reward function R based on the power requirement for target path tracking and the target speed. R=- l 1(P SOC +P BP +P z )+ l 2( ); In the formula, l 1 represents the penalty weighting coefficient. l 2 represents the reward weighting coefficient, P SOC For the state of charge penalty, P BP For charge / discharge power penalty, P z For path tracing, power penalty is required, v r For actual speed, v c This is the command speed.

[0048] It should be noted that dimensional uniformity is required in the calculation. In addition to considering the state of charge and charging / discharging power, the reward function also incorporates the power required for path tracking. When the path tracking deviation is large, additional power is needed for correction (such as large rudder angle maneuvers). In this case, a penalty is applied in the reward function to encourage the DRL to reserve power in advance. Simultaneously, a new speed tracking reward term is added to encourage the actual speed to approach the commanded speed, thereby improving navigation efficiency. This invention incorporates path tracking into the energy management optimization objective during navigation, achieving joint optimization of navigation maneuvers and energy allocation.

[0049] S44. Network Training: A deep Q network (DQN) is used for training. After initializing the network, iterative training is performed, and training termination conditions are set to complete the network training.

[0050] Understandably, in DRL training, the environment model includes not only the original energy system model but also a power demand model from path tracing and a ship motion model. The trained intelligent agent can then output power allocation commands (DE) in real time. t and BP t ).

[0051] S45. Energy management strategy acquisition: Acquire state variables through sensors and path tracking data, input them into the network after normalization to obtain actions, perform constraint checks on the actions and correct out-of-bounds actions, and control the generator and battery to output power according to the corrected actions.

[0052] Reinforcement learning is a branch of machine learning. In this embodiment, the agent learns a policy set by interacting with the path tracing system and receiving feedback. The goal is to obtain the maximum cumulative reward based on the power demand during the path tracing process. The ultimate goal of the energy management strategy is to improve the yacht's range by managing the hybrid power system, based on achieving the target path tracing.

[0053] It should be noted that the path tracing and return to the target path of the present invention can also be applied to the path planning of yacht navigation, and is not limited to path tracing. It is only necessary to set the target destination of the navigation as a preset path.

[0054] This embodiment also provides a new energy intelligent yacht power system, including a photovoltaic cell pack, a generator set, a power battery, an electric motor, and a load.

[0055] The photovoltaic cell array is used to convert solar energy into electrical energy, which is connected to a DC bus via a DC / DC converter to directly power the load or charge the battery. The generator set is used to provide stable, high-power power output, and the output AC power is rectified by an AC / DC converter and then connected to a DC bus. The power battery is used for energy storage and buffering, balancing supply and demand fluctuations, providing short-term high power output, and is connected to a DC bus via a bidirectional DC / DC converter to support charging and discharging. The electric motor and load are used to convert electrical energy into mechanical energy to drive the yacht propulsion system. It should be noted that during operation, the DC bus outputs 380V AC power to the electric motor through the DC / AC inverter. The load includes both powered and non-powered loads. Powered loads include the propulsion system, while non-powered loads include communication and control systems.

[0056] Understandably, in the aforementioned power system, each component performs its function as follows: the AC / DC converter rectifies the alternating current from the generator into direct current; the bidirectional DC / DC converter regulates the battery charging and discharging power; and the DC / AC inverter provides alternating current to the motor and low-voltage loads. In this embodiment, the generator can be a fuel cell, a gas turbine, or a diesel engine, etc., and is not limited thereto.

[0057] It should be noted that this embodiment sets up the above-mentioned power system composition to provide the basis for multi-energy interface management in the energy management optimization method, and realizes the energy distribution of power output.

[0058] This invention first introduces the current energy state of the yacht as a reference point selection interval during path tracing, thereby optimizing the parameters of the rudder angle for path tracing to accommodate the energy state considerations in the path tracing process of energy management in this embodiment. Then, the power demand during path tracing is incorporated into the energy management of the hybrid yacht based on DRL, enabling the energy management system to "anticipate" future sailing needs and make more reasonable energy allocation decisions. By achieving closed-loop linkage of path planning, rudder angle control, power calculation, and energy allocation, a two-way synergy between "sailing mission and energy response" is realized, solving the contradiction between energy consumption and accuracy of new energy yachts in complex environments. This achieves dynamic synergy between path tracing and energy management, providing a high-efficiency power control paradigm for intelligent yachts.

[0059] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. An energy management optimization method for a new energy intelligent yacht power system, characterized in that: Includes the following steps: Obtain the tracking path: Generate a tracking path by selecting target points based on the yacht's current position and the target path; Obtain the tracking rudder angle: Based on the control fineness requirements and energy status, select the spacing of the reference points, select a set of rudder angle control reference points starting from the starting point of the tracking path, and calculate the servo angle; Obtain power requirements: Calculate the yacht's power requirements based on the yacht's steering gear angle and speed; Yacht Energy Management: Energy management strategies are derived from power requirements under yacht operating conditions using deep reinforcement learning; The calculated servo rotation angle includes: Based on the yacht's current energy state, the reference point selection distance Δ is calculated. s : ； In the formula, S 0 is the reference point spacing. γ The energy influence coefficient. E S Energy tension coefficient; Based on the deviation between the yacht's current position and the reference point, it is converted into the deflection angle of the rudder. Then, based on the current operating data, the deflection angle is corrected for errors to obtain the final tracking rudder angle.

2. The energy management optimization method for a new energy intelligent yacht power system according to claim 1, characterized in that: The error correction of the deflection angle includes: A dataset is built using historical navigation data. After preprocessing the collected raw data, features for prediction are extracted from the raw data using the correlation coefficient method. Constructing a BP neural network model: Set the number of nodes in the input layer according to the selected prediction features, set the output layer to compensate for bias, and allocate the collected data into training set, validation set and test set to complete the model training; Based on the predictive features selected through model training, corresponding sensors are set up to collect operating condition data in real time. The collected data is then input into the machine learning model to predict and output the rudder angle to compensate for the deviation, thereby correcting the error in the deflection angle.

3. The energy management optimization method for a new energy intelligent yacht power system according to claim 1, characterized in that: The process of obtaining the tracking path includes the following steps: Selecting target points: Sample a set of candidate target points from the target path, filter the candidate target points by traversing the candidate target points to select the target points that meet the threshold range of direction difference and curvature difference, and take the point closest to the current position as the final target point; Generate initial tracking path: Based on the ship's current position and the selected target point, the initial tracking path of the yacht is randomly generated using B-spline control points; Searching for control points: The control points of the B-spline curve are encoded, and the optimal control point configuration is searched through the evolutionary process of a genetic algorithm to achieve the optimal tracking path planning goal.

4. The energy management optimization method for a new energy intelligent yacht power system according to claim 3, characterized in that: The generation of the initial tracking path also includes the following steps: Let the time domain be The number of control points is n The B-spline curve locus is: ； In the formula, For the spline curve locus, t ∈ ; These are the control points for the B-spline curve. For the first i indivual k B-order spline basis functions, basis functions It is constructed using the de Boer-Cox recurrence relation, and its expression is: ； In the formula, U 1. U 2 is a basis function The coefficient; Introducing B-spline curves into the representation of ship paths, a set of control point coordinates can be expressed as: ； The path expression generated by this set of control points is: ； In the formula, X ( t ) indicates that a point on the path is at x Changes in position on the axis. Y ( t ) indicates that a point on the path is at y The positional changes on the axis.

5. The energy management optimization method for a new energy intelligent yacht power system according to claim 4, characterized in that: The search for control points also includes the following steps: Encode the control points and solve for the control points ( , )coordinate: ； In the formula, ( , ( ) is the starting point of the path s and the end point f Connection sf The coordinates of the division points, Starting point s and the end point f The distance between them rand are random numbers and rand ∈[-0.5,0.5], It is a straight line sf The slope; Establish the fitness function: Based on the actual needs in the error correction process, set constraints to establish the fitness function. The formula for the fitness function is: ； In the formula, , , These are the weighting coefficients. The path distance. For path smoothness, The path direction should match the objective function; Genetic Iterative Optimization: Minimize the fitness function through iterative optimization to obtain the optimal path.

6. The energy management optimization method for a new energy intelligent yacht power system according to claim 1, characterized in that: The energy tension coefficient E S The calculation formula is: ； In the formula, SOC The battery is in its state of charge. P pv The average power of photovoltaic power. k for SOC Sensitivity coefficient α This is the photovoltaic weighting coefficient.

7. The energy management optimization method for a new energy intelligent yacht power system according to claim 1, characterized in that: The process of obtaining power requirements includes the following steps: Propulsion power is obtained based on the ship's motion. and dynamic compensation power Δ P d The calculation formula is: ； In the formula, ρ For the density of water, C d This is the hull drag coefficient. A This represents the projected area of ​​the ship's hull in contact with the water. v c At the current speed, C w The coefficient of water resistance. θ The pitch angle; ； In the formula, a For acceleration; Yacht steering power is obtained from the yacht's rudder angle. The calculation formula is: ； In the formula, The torque coefficient, δ For servo motor turning angle, ω The angular velocity of the rudder blade. η For the efficiency of the hydraulic system, k l This is the steering loss coefficient; Based on propulsion power, dynamic compensation power, and yacht steering power, calculate the yacht's total power requirement: P z = +D P d + ; In the formula, P z This represents the total power requirements of the yacht.

8. The energy management optimization method for a new energy intelligent yacht power system according to claim 1, characterized in that: The yacht energy management includes the following steps: Define the state space: Construct the state space S of the hybrid yacht by including the yacht's state detection variables and all state variables required for path tracing. t : S t =[PV t ,SOC t ,P zt , v ct ]; In the formula, PV t Let be the photovoltaic power at time t, and SOC. t Let P be the battery state of charge at time t. zt Let t be the power demand. v ct The speed is the commanded speed at time t; Setting the Action Space: Based on the generator and power battery, setting the action space A of the hybrid yacht. t : TO t =[OF t BP t ]; In the formula, DE t For the generator output, BP t The power used to charge or discharge the power battery; The constraints include: ； ≤DE t ≤min( DE max , + ); In the formula, BP max Δ is the maximum allowable discharge power of the battery. P d For dynamic power compensation, DE max This is the maximum allowable output power of the generator. For yacht propulsion power, For yacht steering power; Set the reward function: Construct a reward function R based on the power requirements of the target path tracking and the target speed. R=- λ 1(P SOC +P BP +P z )+ λ 2( ); In the formula, λ 1 represents the penalty weighting coefficient. λ 2 represents the reward weighting coefficient, P SOC For the state of charge penalty, P BP For charge / discharge power penalty, P z For path tracing, power penalty is required, v r For actual speed, v c This is the command speed.

9. The energy management optimization method for a new energy intelligent yacht power system according to claim 8, characterized in that: The yacht energy management also includes the following steps: Network training: A deep Q-network is used for training. After initializing the network, iterative training is performed, and training termination conditions are set to complete the network training. Energy management strategy acquisition: State variables are acquired through sensors and path tracking data, and after normalization, they are input into the network to acquire actions. Constraint checks are performed on the actions and out-of-bounds actions are corrected. The generator and battery are controlled to output power according to the corrected actions.

10. A new energy intelligent yacht power system, implementing the energy management optimization method for a new energy intelligent yacht power system as described in any one of claims 1-9, characterized in that: This includes photovoltaic cell modules, generator sets, power batteries, electric motors, and loads; The photovoltaic cell array is used to convert solar energy into electrical energy, which is connected to a DC bus via a DC / DC converter to directly power the load or charge the battery. The generator set is used to provide stable, high-power power output, and the output AC power is rectified by an AC / DC converter and then connected to a DC bus. The power battery is used for energy storage and buffering, balancing supply and demand fluctuations, providing short-term high power output, and is connected to a DC bus via a bidirectional DC / DC converter to support charging and discharging. The electric motor and load are used to convert electrical energy into mechanical energy to drive the yacht's propulsion system.

Images