Full-revolving propeller simulation control method and system based on empirical formula prior

Through the azimuth thruster simulation control method based on empirical formula, combined with the fluid inertia force, viscous force and rudder force models, the Adam optimization algorithm is used to construct the azimuth thruster simulation control system, which solves the problems of high cost and insufficient accuracy of azimuth thruster simulation training, and realizes the construction of low-cost and high-precision ship maneuverability model.

CN120595799APending Publication Date: 2025-09-05SHANGHAI SHIP & SHIPPING RES INST CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510712621.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-29
Publication Date
2025-09-05

AI Technical Summary

Technical Problem

The existing azimuth thruster simulation training system is expensive and lacks accuracy, making it difficult to build a low-cost, high-precision ship maneuverability model.

Method used

A method based on empirical formula is adopted to obtain ship operating parameters and propeller parameters. Combined with the fluid inertia force, viscous force and rudder force models, the Adam optimization algorithm is used to construct an azimuth thruster simulation control system, which reduces the experimental cost of the maneuverability coefficient and improves the model accuracy.

Benefits of technology

It realizes low-cost, high-precision full-rotation thruster simulation training, provides a scenario close to real ship operation, is suitable for large-scale teaching and demonstration, and reduces the experimental cost of maneuverability coefficient fitting and the cost of setting up the simulation system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120595799A_ABST
    Figure CN120595799A_ABST
Patent Text Reader

Abstract

The invention relates to a full-revolving thruster simulation control method and system based on empirical formula prior, and the method comprises the steps: firstly obtaining ship operation parameters, thruster parameters and rudder parameters, correcting the thrust reduction coefficient of a propeller, and calculating the force and moment acting on a bare ship body, the propeller and a rudder through an empirical formula and a specific calculation method, the forward acceleration, the lateral acceleration and the bow turning angle acceleration at the current moment are calculated through a ship motion equation; time integral operation is conducted on the forward acceleration, the lateral acceleration and the bow turning angular acceleration at the current moment to obtain the forward speed, the lateral speed and the bow turning angular speed at the next moment of the current moment, ship state data at the next moment of the current moment are calculated through a Runge-Kutta method, and then a target function is constructed; adopting an Adam optimization algorithm to calculate and obtain an optimal target optimization variable; and constructing a full-revolving propeller ship control motion model based on the optimal target optimization variable, and realizing training of ship drivers.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of ship maneuvering simulation and control in ship and ocean engineering, and in particular to a method and system for simulating and controlling an azimuth propeller based on empirical formula priors. Background Art

[0002] Azimuth thrusters can rotate freely and generate thrust in any direction, simultaneously propelling and maneuvering the vessel. They are widely used on specialized vessels such as offshore engineering vessels, icebreakers, and research vessels. Azimuth thrusters have a specific operating method and logic. The thruster handle can be used to control thrust forward and backward, as well as to rotate to control thrust direction. Unlike traditional thrusters, these thrusters are more difficult to operate and require training for ship operators.

[0003] The ship driving training program mainly includes simulation training and actual ship training. The actual ship training is very expensive and it is difficult to conduct operation training in emergency situations. There are also fewer opportunities to correct incorrect operations. Traditional ship simulation training requires a large venue and expensive supporting hardware and software equipment. It is not suitable for preliminary training on the operation of full-turn propellers and is not conducive to large-scale demonstration and teaching. It is worth noting that the core reliance of simulation training is the accuracy of the ship maneuverability model, and the calculation of the existing ship maneuverability model requires maneuverability tests to obtain many coefficients in the model. The test is cumbersome and costly. In other words, the existing design schemes either completely use empirical formulas to calculate the maneuverability coefficients, or obtain the maneuverability coefficients through physical tests. The former has a slightly lower accuracy of the maneuverability model, and the latter is expensive, making it difficult to establish a low-cost and simple-to-build simulation test scenario.

[0004] Therefore, developing a low-cost, experiential full-rotation thruster simulation training system requires breaking through the technical bottleneck of traditional model construction and achieving both accuracy and economy, which has become an urgent need in the industry. Summary of the Invention

[0005] To address the current problems in ship simulation training, where design solutions either rely solely on empirical formulas to calculate maneuverability coefficients or obtain maneuverability coefficients through physical testing, resulting in poor maneuverability model accuracy and high costs, the present invention provides a method for simulating and controlling an azimuth propeller based on empirical formulas. This method eliminates the need to obtain numerous coefficients in the ship maneuverability model, simplifies the test, and significantly reduces the cost of fitting the maneuverability coefficients. It more accurately simulates the maneuvering motion of a ship equipped with an azimuth propeller, providing a scenario closer to actual ship maneuvering for crew training. The present invention also relates to a azimuth propeller simulation control system based on empirical formulas.

[0006] The technical solutions of the present invention are as follows:

[0007] A method for simulating and controlling an azimuth thruster based on an empirical formula prior is characterized by comprising the following steps:

[0008] Parameter acquisition step: acquiring ship operation parameters, propeller parameters and rudder parameters, and establishing a hull coordinate system with the ship's center of gravity as the origin, the direction of the ship's bow as the longitudinal axis, the direction of the ship's starboard side as the transverse axis, and the vertically downward direction of the ship's bottom as the vertical axis; the ship operation parameters include ship mass, ship length, ship width, bow draft, stern draft, hull wetted area, ship's forward speed, lateral speed and bow angular velocity; calculating the combined speed of the ship's actual motion based on the forward speed and lateral speed, and calculating the average draft based on the bow draft and stern draft; the propeller parameters include the total thrust of the propeller, the thrust direction angle, the basic deduction coefficient, the direction correction coefficient, the propeller's thrust deduction coefficient and the position coordinates of the propeller in the hull coordinate system; the rudder parameters include the rudder's positive pressure, rudder angle, rudder resistance deduction coefficient, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's positive pressure application point to the ship's center of gravity;

[0009] Calculation steps for fluid inertia force and moment: Based on the ship's mass, length, width and average draft, the longitudinal added mass and transverse added mass are calculated using empirical formulas; based on the ship's length and width, the vertical added moment of inertia is calculated using empirical formulas; based on the longitudinal added mass, transverse added mass, forward speed, lateral speed and bow angular velocity, the longitudinal fluid inertia force and transverse fluid inertia force are calculated respectively, and the bow fluid inertia moment is calculated based on the vertical added moment of inertia, the longitudinal added mass, transverse added mass, forward speed, lateral speed and bow angular velocity;

[0010] Calculation steps for fluid viscous force and torque: Calculate the dimensionless forward velocity based on the resultant velocity and forward velocity; calculate the dimensionless lateral velocity based on the resultant velocity and lateral velocity; and calculate the dimensionless bow angular velocity based on the resultant velocity, ship length, and bow angular velocity; calculate the straight-line resistance coefficient based on the hull wetted area, ship length, average draft, and dimensionless forward velocity, and then calculate the fluid dynamic derivative based on the ship length, ship width, and average draft using an empirical formula; based on the straight-line resistance coefficient, fluid dynamic derivative, dimensionless forward velocity, dimensionless lateral velocity, and dimensionless bow angular velocity, the dimensionless longitudinal viscous force, dimensionless transverse viscous force, and dimensionless bow viscous moment are calculated using the dimensionless hull viscous force nonlinear model;

[0011] Calculation steps for the forces and moments acting on the bare hull: Calculate the longitudinal force acting on the bare hull based on the longitudinal fluid inertia force and the dimensionless longitudinal viscous force; calculate the transverse force acting on the bare hull based on the transverse fluid inertia force and the dimensionless transverse viscous force; and calculate the rotational moment acting on the bare hull about the center of gravity of the ship based on the bow fluid inertia moment and the dimensionless bow viscous moment;

[0012] Calculation steps for the force and torque acting on the propeller: construct a linear correction model based on the thrust direction angle, basic derating coefficient and directional correction coefficient of the propeller, and use the linear correction model to correct the thrust derating coefficient of the propeller to obtain the corrected thrust derating coefficient; calculate the longitudinal thrust of the propeller and the transverse thrust of the propeller according to the total thrust of the propeller, the thrust direction angle of the propeller, the corrected thrust derating coefficient, the basic derating coefficient and the directional correction coefficient, and then obtain the longitudinal force acting on the propeller and the transverse force acting on the propeller; then calculate the rotational torque acting on the propeller around the center of gravity of the ship according to the position coordinates of the propeller in the hull coordinate system, the longitudinal thrust of the propeller and the transverse thrust of the propeller;

[0013] Calculation steps for the forces and moments acting on the rudder: Calculate the longitudinal force acting on the rudder based on the rudder's normal pressure, rudder angle, and rudder resistance reduction coefficient. Calculate the lateral force acting on the rudder based on the rudder's normal pressure and rudder angle. Calculate the rotational moment acting on the rudder about the ship's center of gravity based on the rudder's normal pressure, rudder angle, the distance from the rudder's point of application of the lateral force to the ship's center of gravity, and the distance from the rudder's point of application of the normal pressure to the ship's center of gravity.

[0014] Ship status data calculation steps: Based on the longitudinal force, lateral force and rotational moment about the ship's center of gravity acting on the bare hull at the current moment, the longitudinal force, lateral force and rotational moment about the ship's center of gravity acting on the propeller, and the longitudinal force, lateral force and rotational moment about the ship's center of gravity acting on the rudder, and using the ship motion equation to calculate the forward acceleration, lateral acceleration and bow angular acceleration at the current moment; then, time integration operations are performed on the forward acceleration, lateral acceleration and bow angular acceleration at the current moment to obtain the forward velocity, lateral velocity and bow angular velocity at the next moment of the current moment; based on the forward velocity, lateral velocity and bow angular velocity at the next moment of the current moment, the Runge-Kutta method is used to calculate the ship status data at the next moment of the current moment;

[0015] The steps for constructing the maneuvering motion model of a ship with an azimuth thruster are as follows: the basic deduction coefficient, direction correction coefficient, straight-line resistance coefficient and fluid dynamic derivative are used as target optimization variables, and the objective function is constructed based on the target optimization variables, the ship status data at the next moment of the current moment and the actual ship status data. The Adam optimization algorithm is used to calculate the optimal solution of the objective function, and then the optimal target optimization variables are obtained; the maneuvering motion model of a ship with an azimuth thruster is constructed based on the optimal target optimization variables to simulate the maneuvering characteristics of the azimuth thruster.

[0016] Preferably, in the step of calculating the fluid viscous force and torque, the fluid dynamic derivative is calculated based on the ship length, ship width and average draft using an empirical formula, specifically including:

[0017] The empirical formulas include the Inoue formula, the Matsumoto formula and the ship fluid dynamic derivative regression formula. The linear fluid dynamic derivative is calculated based on the ship length, ship width and average draft using the Inoue formula; the second-order coupling derivative is calculated based on the lateral added mass; the first group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the Matsumoto formula; the second group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the ship fluid dynamic derivative regression formula; and the fluid dynamic derivative is generated based on the linear fluid dynamic derivative, the second-order coupling derivative, the first group of nonlinear fluid dynamic derivatives and the second group of nonlinear fluid dynamic derivatives.

[0018] Preferably, the linear fluid dynamic derivatives include the derivative of lateral force to lateral velocity, the derivative of lateral force to bow angular velocity, the derivative of bow moment to lateral velocity, and the derivative of bow moment to bow angular velocity; the first group of nonlinear fluid dynamic derivatives include second-order lateral velocity derivatives and second-order bow angular velocity derivatives, and the second group of nonlinear fluid dynamic derivatives include first high-order nonlinear derivatives, second high-order nonlinear derivatives, first mixed absolute value derivatives, second mixed absolute value derivatives, first high-order mixed derivatives, and second high-order mixed derivatives.

[0019] Preferably, in the ship status data calculation step, the ship status data includes the ship's position and heading.

[0020] Preferably, in the step of constructing the full-turn propeller ship maneuvering motion model, the Adam optimization algorithm adaptively adjusts the learning rate of each parameter in the target optimization variable through the first-order moment and the second-order moment to avoid the complexity of manual parameter adjustment.

[0021] A simulation control system for an azimuth thruster based on empirical formula priors is characterized by comprising a parameter acquisition module, a fluid inertia force and torque calculation module, a fluid viscous force and torque calculation module, a force and torque calculation module acting on the bare hull, a force and torque calculation module acting on the propeller, a force and torque calculation module acting on the rudder, a ship state data calculation module, and an azimuth thruster ship maneuvering motion model construction module, which are connected in sequence.

[0022] The parameter acquisition module acquires ship operation parameters, propeller parameters and rudder parameters, and establishes a hull coordinate system with the ship's center of gravity as the origin, the direction of the ship's bow as the longitudinal axis, the direction of the ship's starboard side as the transverse axis, and the vertically downward direction of the ship's bottom as the vertical axis; the ship operation parameters include ship mass, ship length, ship width, bow draft, stern draft, hull wetted area, ship's forward speed, lateral speed and bow angular velocity; the actual motion resultant speed of the ship is calculated based on the forward speed and lateral speed, and the average draft is calculated based on the bow draft and stern draft; the propeller parameters include the total thrust of the propeller, the thrust direction angle, the basic deduction coefficient, the direction correction coefficient, the propeller thrust deduction coefficient and the position coordinates of the propeller in the hull coordinate system; the rudder parameters include the rudder's positive pressure, rudder angle, rudder resistance deduction coefficient, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's positive pressure application point to the ship's center of gravity;

[0023] The fluid inertia force and moment calculation module calculates the longitudinal added mass and the transverse added mass based on the ship mass, ship length, ship width and average draft using empirical formulas; calculates the vertical added moment of inertia based on the ship length and ship width using empirical formulas; calculates the longitudinal fluid inertia force and the transverse fluid inertia force based on the longitudinal added mass, the transverse added mass, the forward speed, the lateral speed and the bow angular velocity, and calculates the bow fluid inertia moment based on the vertical added moment of inertia, the longitudinal added mass, the transverse added mass, the forward speed, the lateral speed and the bow angular velocity;

[0024] The fluid viscosity force and torque calculation module calculates the dimensionless forward velocity according to the resultant velocity and the forward velocity; calculates the dimensionless lateral velocity according to the resultant velocity and the lateral velocity; and calculates the dimensionless bow angular velocity according to the resultant velocity, the ship length, and the bow angular velocity; calculates the straight-line resistance coefficient according to the hull wetted area, the ship length, the average draft, and the dimensionless forward velocity, and then calculates the fluid dynamic derivative based on the ship length, the ship width, and the average draft using an empirical formula; and calculates the dimensionless longitudinal viscous force, the dimensionless transverse viscous force, and the dimensionless bow viscous moment based on the straight-line resistance coefficient, the fluid dynamic derivative, the dimensionless forward velocity, the dimensionless lateral velocity, and the dimensionless bow angular velocity using a dimensionless hull viscosity nonlinear model;

[0025] The force and moment calculation module acting on the bare hull calculates the longitudinal force acting on the bare hull based on the longitudinal fluid inertia force and the dimensionless longitudinal viscous force; calculates the lateral force acting on the bare hull based on the lateral fluid inertia force and the dimensionless lateral viscous force; and calculates the rotational moment acting on the bare hull around the center of gravity of the ship based on the bow fluid inertia moment and the dimensionless bow viscous moment;

[0026] The force and torque calculation module acting on the propeller constructs a linear correction model based on the thrust direction angle, basic deduction coefficient and direction correction coefficient of the propeller, and uses the linear correction model to correct the thrust deduction coefficient of the propeller to obtain a corrected thrust deduction coefficient; according to the total thrust of the propeller, the thrust direction angle of the propeller, the corrected thrust deduction coefficient, the basic deduction coefficient and the direction correction coefficient, the longitudinal thrust of the propeller and the transverse thrust of the propeller are calculated respectively, thereby obtaining the longitudinal force acting on the propeller and the transverse force acting on the propeller; and then, according to the position coordinates of the propeller in the hull coordinate system, the longitudinal thrust of the propeller and the transverse thrust of the propeller, the rotational torque around the center of gravity of the ship is calculated;

[0027] The rudder force and torque calculation module calculates the longitudinal force acting on the rudder based on the rudder's normal pressure, rudder angle, and rudder resistance deduction coefficient, calculates the lateral force acting on the rudder based on the rudder's normal pressure and rudder angle, and calculates the rotational torque acting on the rudder about the ship's center of gravity based on the rudder's normal pressure, rudder angle, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's normal pressure application point to the ship's center of gravity;

[0028] The ship state data calculation module calculates the forward acceleration, lateral acceleration and bow angular acceleration at the current moment based on the longitudinal force, lateral force and rotational torque about the ship's center of gravity acting on the bare hull, the longitudinal force, lateral force and rotational torque about the ship's center of gravity acting on the propeller, and the longitudinal force, lateral force and rotational torque about the ship's center of gravity acting on the rudder at the current moment, and uses the ship motion equation to calculate; then, time integration operations are performed on the forward acceleration, lateral acceleration and bow angular acceleration at the current moment to obtain the forward velocity, lateral velocity and bow angular velocity at the next moment of the current moment; based on the forward velocity, lateral velocity and bow angular velocity at the next moment of the current moment, the Runge-Kutta method is used to calculate the ship state data at the next moment of the current moment;

[0029] The module for constructing the ship maneuvering motion model of an azimuth thruster uses the basic deduction coefficient, the directional correction coefficient, the straight-line resistance coefficient, and the fluid dynamic derivative as target optimization variables, constructs an objective function based on the target optimization variables, the ship status data at the next moment of the current moment, and the actual ship status data, and uses the Adam optimization algorithm to calculate the optimal solution of the objective function, thereby obtaining the optimal target optimization variables; the module for constructing the ship maneuvering motion model of an azimuth thruster is constructed based on the optimal target optimization variables to simulate the maneuvering characteristics of the azimuth thruster.

[0030] Preferably, in the fluid viscous force and torque calculation module, the fluid dynamic derivative is calculated based on the ship length, ship width and average draft using an empirical formula, specifically including:

[0031] The empirical formulas include the Inoue formula, the Matsumoto formula and the ship fluid dynamic derivative regression formula. The linear fluid dynamic derivative is calculated based on the ship length, ship width and average draft using the Inoue formula; the second-order coupling derivative is calculated based on the lateral added mass; the first group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the Matsumoto formula; the second group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the ship fluid dynamic derivative regression formula; and the fluid dynamic derivative is generated based on the linear fluid dynamic derivative, the second-order coupling derivative, the first group of nonlinear fluid dynamic derivatives and the second group of nonlinear fluid dynamic derivatives.

[0032] Preferably, the linear fluid dynamic derivatives include the derivative of lateral force to lateral velocity, the derivative of lateral force to bow angular velocity, the derivative of bow moment to lateral velocity, and the derivative of bow moment to bow angular velocity; the first group of nonlinear fluid dynamic derivatives include second-order lateral velocity derivatives and second-order bow angular velocity derivatives, and the second group of nonlinear fluid dynamic derivatives include first high-order nonlinear derivatives, second high-order nonlinear derivatives, first mixed absolute value derivatives, second mixed absolute value derivatives, first high-order mixed derivatives, and second high-order mixed derivatives.

[0033] Preferably, the ship status data includes the ship's position and heading.

[0034] Preferably, the Adam optimization algorithm adaptively adjusts the learning rate of each parameter in the target optimization variable through the first-order moment and the second-order moment to avoid the complexity of manual parameter adjustment.

[0035] The technical effects of the present invention are as follows:

[0036] The present invention provides a method for simulating and controlling a full-turn propeller based on an empirical formula prior. First, the ship's operating parameters, propeller parameters and rudder parameters are obtained, and the longitudinal fluid inertia force, the lateral fluid inertia force and the bow fluid inertia moment are calculated respectively by using empirical formulas such as the ship's fluid dynamics derivative regression formula combined with a specific calculation method. This method can accurately estimate the "virtual mass" effect of the fluid around the ship on the motion and improve the dynamic response accuracy of the model. Then, the dimensionless forward velocity, lateral velocity and bow angular velocity are obtained through dimensionless processing to eliminate unit differences and improve the versatility and stability of the model. The dimensionless longitudinal viscous force, dimensionless lateral viscous force and dimensionless bow angular velocity are calculated by using empirical formulas combined with a specific calculation method. The dimensional bow viscous moment is calculated, and then the longitudinal force, lateral force and rotational moment around the center of gravity of the ship acting on the bare hull are calculated, which enhances the realism and prediction accuracy of the model; then, a linear correction model is constructed based on the thrust direction angle, basic deduction coefficient and direction correction coefficient of the propeller, and the thrust deduction coefficient of the propeller is corrected by the linear correction model to obtain the corrected thrust deduction coefficient. By correcting the thrust deduction coefficient of the propeller, the influence of the actual propulsion efficiency changing with the direction angle is effectively considered, and the accuracy of the subsequent fluid dynamic model of the full-rotating propeller is improved. The maneuvering movement of the ship equipped with the full-rotating propeller is simulated more accurately, providing a scene closer to the actual ship operation for crew training, which is conducive to large-scale teaching and exhibition The results show that the problem of inaccurate simulation caused by the change of propeller angle is solved; the longitudinal force, lateral force and turning moment around the center of gravity of the ship acting on the propeller are calculated by a specific calculation method, and the authenticity of the control simulation is improved by considering the influence of the propeller installation position on the bow turning moment; and the longitudinal force, lateral force and turning moment around the center of gravity of the ship of the rudder are calculated by a specific calculation method, and the motion characteristics of the ship can be accurately simulated by combining the characteristics of the steering gear. By considering the influence of the rudder force on the bow turning performance, the control ability of the rudder over the ship's rotation is improved; then the forward acceleration, lateral acceleration and bow angular acceleration at the current moment are calculated using the ship motion equation, and the time integration combined with the Runge-Kutta method is used to calculate the current moment. The ship status data at the next moment is obtained, which effectively improves the stability and accuracy of the numerical solution and makes the simulation process closer to the real physical behavior; finally, the basic deduction coefficient, direction correction coefficient, straight-line resistance coefficient and fluid dynamic derivative are used as target optimization variables to construct the objective function, and the Adam optimization algorithm is used to calculate the optimal solution of the objective function, and then the optimal target optimization variables are obtained. By using the Adam optimization algorithm to automatically calibrate the model parameters, the model can adapt to different ship types and environmental conditions, and can be more in line with the measured data, which significantly improves the model accuracy. Based on the optimal target optimization variables, a full-rotation thruster ship maneuvering motion model is constructed to simulate the maneuvering characteristics of the full-rotation thruster and realize the training of ship drivers.The present invention iteratively optimizes the basic ship parameters a priori and a small amount of experimental data through empirical formulas, comprehensively considers the interaction between fluid inertia force, viscous force, propeller force and rudder force, performs multi-force coupling calculations, and solves the motion equations through the Runge-Kutta method to improve the dynamic response accuracy. It avoids the high cost of traditional physical tests and solves the accuracy problem through angle correction and parameter iteration. Ultimately, it provides a simulation scene close to the real ship for crew training, and comprehensively solves the pain points of the existing technology.

[0037] Furthermore, in the fluid viscous force and torque calculation step, based on the ship length, ship width and average draft, the fluid dynamic derivatives are calculated using empirical formulas, specifically including: empirical formulas include Inoue formula, Matsumoto formula and ship fluid dynamic derivative regression formula (such as Zhou Zhaoming regression formula), based on the ship length, ship width and average draft, the linear fluid dynamic derivative is calculated using the Inoue formula; the second-order coupled derivative is calculated based on the lateral added mass; based on the ship length, ship width and average draft, the first group of nonlinear fluid dynamic derivatives is calculated using the Matsumoto formula; based on the ship length, ship width and average draft, the second group is calculated using the Zhou Zhaoming regression formula Nonlinear fluid dynamic derivatives: Fluid dynamic derivatives are generated based on linear fluid dynamic derivatives, second-order coupled derivatives, the first set of nonlinear fluid dynamic derivatives, and the second set of nonlinear fluid dynamic derivatives. By integrating multiple classical empirical formulas for comprehensive calculation and introducing prior knowledge, realistic ship motion simulation can be achieved by combining only a small amount of ship maneuvering data, without relying on complex CFD simulations or actual ship tests. It has high engineering practicality, greatly reduces the experimental cost of fitting the maneuverability coefficient and the cost of setting up the simulation system, and significantly reduces the cost and cycle of ship motion modeling. It is suitable for the design and optimization of various ship model control systems. In addition, it can more accurately reflect the fluid dynamic characteristics of ships under different navigation conditions, especially under complex motion conditions such as low speed, medium and high speed, and turning, with good adaptability and prediction accuracy; it avoids the model error caused by a single formula, improves the comprehensiveness and reliability of the estimation of viscous force and its torque, and thus provides a more accurate dynamic basis for the modeling and simulation of control systems.

[0038] In the step of constructing the full-turn thruster ship maneuvering motion model, the Adam optimization algorithm adaptively adjusts the learning rate of each parameter in the target optimization variable through the first-order moment and the second-order moment to avoid the complexity of manual parameter adjustment. The algorithm combines momentum (first-order moment estimation) and adaptive learning rate adjustment mechanism (second-order moment estimation), and can dynamically adjust its update step size according to the gradient change characteristics of different parameters during the parameter optimization process, thereby accelerating the convergence speed and improving the optimization stability, reducing the dependence on the experience of the modeler, and improving the automation level of model construction. It is suitable for the rapid modeling needs of different ship types and various navigation conditions.

[0039] The present invention integrates the empirical formula of ship fluid dynamics, the force model of the propeller and rudder, and advanced numerical integration and optimization algorithms to construct a high-precision, high-real-time, and highly adjustable full-rotation propeller ship maneuvering simulation system. By combining the empirical formula to obtain a rough estimate of the ship maneuverability coefficient, prior knowledge is introduced, and then the Adam (Adaptive Moment Estimation) optimization algorithm is used. The prior knowledge introduced based on the empirical formula enables the Adam optimization process to achieve realistic ship motion simulation by combining only a small amount of ship maneuvering data, without relying on complex CFD simulation or actual ship testing. It has high engineering practicality, significantly reduces the cost and cycle of ship motion modeling, and is suitable for the design and optimization of various ship model control systems. It further fits the maneuverability coefficient and achieves realistic ship motion simulation by comprehensively considering multiple factors such as fluid inertia force, viscous force, propeller thrust, rudder force, etc., greatly reducing the experimental cost of fitting the maneuverability coefficient and the cost of building the simulation system.

[0040] The present invention also relates to an empirical formula-based full-rotation propeller simulation control system, which corresponds to the above-mentioned empirical formula-based full-rotation propeller simulation control method, and can be understood as a system for realizing the above-mentioned empirical formula-based full-rotation propeller simulation control method, including a parameter acquisition module, a fluid inertia force and torque calculation module, a fluid viscosity force and torque calculation module, a force and torque calculation module acting on the bare hull, a force and torque calculation module acting on the propeller, and a force and torque calculation module acting on the rudder, which are connected in sequence. , a ship status data calculation module, and an azimuth thruster ship maneuvering motion model construction module. Each module works together to build a high-precision, high-real-time, and highly adjustable azimuth thruster ship maneuvering simulation system by integrating empirical formulas for ship fluid dynamics, force models of the propeller and rudder, and advanced numerical integration and optimization algorithms. By correcting the propeller's thrust reduction coefficient, a fluid dynamic model of the azimuth propeller is established, which more accurately simulates the maneuvering motion of a ship equipped with an azimuth propeller, providing a scenario closer to actual ship maneuvering for crew training, and facilitating large-scale teaching and demonstration. By combining empirical formulas to obtain a rough estimate of the ship's maneuverability coefficient and introducing prior knowledge, the Adam (Adaptive Moment Estimation) optimization algorithm is then used to further fit the maneuverability coefficient in combination with a small amount of ship maneuvering data. By comprehensively considering multiple factors such as fluid inertia, viscosity, propeller thrust, and rudder force, a realistic ship motion simulation is achieved, greatly reducing the experimental cost of fitting the maneuverability coefficient and the cost of building the simulation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1It is a flow chart of the present invention's method for simulating and controlling an azimuth propeller based on an empirical formula.

[0042] Figure 2 It is an optimal flow chart of the present invention's method for simulating and controlling an azimuth propeller based on an empirical formula. DETAILED DESCRIPTION

[0043] The present invention will be described below with reference to the accompanying drawings.

[0044] The present invention relates to a method for simulating and controlling an azimuth thruster based on an empirical formula. The flow chart of the method is shown in FIG. Figure 1 As shown, the following steps are included in sequence:

[0045] 1. Parameter acquisition step: obtain ship operating parameters, propeller parameters and rudder parameters, and establish a hull coordinate system with the ship's center of gravity as the origin, the direction of the ship's bow as the longitudinal axis, the direction of the ship's starboard side as the transverse axis, and the vertically downward direction of the ship's bottom as the vertical axis; the ship operating parameters include ship mass, ship length, ship width, bow draft, stern draft, hull wetted area, ship's forward speed, lateral speed and bow angular velocity; calculate the actual motion resultant speed of the ship based on the forward speed and lateral speed, and calculate the average draft based on the bow draft and stern draft; the propeller parameters include the total thrust of the propeller, thrust direction angle, basic deduction coefficient, direction correction coefficient, propeller thrust deduction coefficient and the position coordinates of the propeller in the hull coordinate system; the rudder parameters include the rudder's positive pressure, rudder angle, rudder resistance deduction coefficient, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's positive pressure force application point to the ship's center of gravity.

[0046] 2. Calculation steps of fluid inertia force and moment: Based on the ship mass, length, width and average draft, the longitudinal added mass and the transverse added mass are calculated respectively using the ship fluid dynamics derivative regression formula in the empirical formula (such as Zhou Zhaoming's regression formula, which is the formula obtained by Zhou Zhaoming through multivariate regression analysis of the Yuanliang map); based on the ship length and width, the vertical additional moment of inertia is calculated using the ship fluid dynamics derivative regression formula; based on the longitudinal added mass, transverse added mass, forward speed, lateral speed and bow angular velocity, the longitudinal fluid inertia force and the transverse fluid inertia force are calculated respectively, and the bow fluid inertia moment is calculated based on the vertical additional moment of inertia, the longitudinal added mass, the transverse added mass, forward speed, lateral speed and bow angular velocity.

[0047] Specifically, the motion of a ship can be approximately regarded as a planar motion from the perspective of open-loop or closed-loop maneuvering. Ignoring the coupling between forward, lateral, and bow motions and heave, roll, and pitch motions, the ship motion equation is expressed as follows:

[0048]

[0049] Formula (1) is the MMG model, where m is the mass of the ship, I zz is the moment of inertia of the ship, u, v, r are the forward speed, lateral speed and bow angular velocity of the ship respectively, are their derivatives respectively. X, Y, and N are the longitudinal force, transverse force, and moment around the center of gravity of the ship. Since N is the moment measured relative to the center of the ship, it is necessary to correct N to the center of gravity. The correction term is Y x C , x C It is the x-axis coordinate value of the ship's center of gravity in the hull coordinate system.

[0050] The specific expressions of each force are shown in the following formula:

[0051]

[0052] in, are the longitudinal force, transverse force and rotational moment around the center of gravity of the ship acting on the bare hull, respectively, X P ,Y P ,N P are the longitudinal force, lateral force and rotational moment around the center of gravity of the ship acting on the propeller, respectively. R ,Y R ,N R are the longitudinal force, lateral force and rotational moment about the center of gravity of the ship acting on the rudder respectively.

[0053] The fluid dynamics and moments acting on the bare hull (i.e., the longitudinal force, transverse force, and rotational moment around the center of gravity of the ship) are decomposed into inertial fluid dynamics and moments and viscous fluid dynamics and moments according to their properties, and then rewritten as shown below:

[0054]

[0055] Inertial fluid dynamics and moments, i.e. longitudinal fluid inertial force X l , lateral fluid inertia force Y l and bow fluid inertia moment N l , its decomposition formula is shown as follows:

[0056]

[0057] Among them, m x ,m y ,J zz are the longitudinal additional mass, the transverse additional mass and the vertical additional moment of inertia (i.e. longitudinal additional mass, transverse additional mass and vertical additional moment of inertia), α xis the additional mass m y The x-coordinate value of the center of gravity is approximately 0, so:

[0058]

[0059] For the longitudinal additional mass, lateral additional mass and vertical additional moment of inertia, that is, m x ,m y ,J zz The calculation adopts the ship hydrodynamic derivative regression formula, such as Zhou Zhaoming's regression formula. Based on the ship mass, length, width and average draft, the longitudinal additional mass and the transverse additional mass are calculated respectively using Zhou Zhaoming's regression formula. Based on the ship length and width, the vertical additional moment of inertia is calculated using Zhou Zhaoming's regression formula. The specific formula is shown below:

[0060]

[0061] Among them, L is the length of the ship, B is the width of the ship, d is the average draft, C is the b is the square coefficient.

[0062] 3. Calculation steps of fluid viscous force and torque: Calculate the dimensionless forward velocity based on the resultant velocity and forward velocity; calculate the dimensionless lateral velocity based on the resultant velocity and lateral velocity; and calculate the dimensionless bow angular velocity based on the resultant velocity, ship length and bow angular velocity; calculate the straight-line resistance coefficient based on the hull wetted area, ship length, average draft and dimensionless forward velocity, and then calculate the fluid dynamic derivative based on the ship length, ship width and average draft using an empirical formula; based on the straight-line resistance coefficient, fluid dynamic derivative, dimensionless forward velocity, dimensionless lateral velocity and dimensionless bow angular velocity, the dimensionless longitudinal viscous force, dimensionless transverse viscous force and dimensionless bow viscous moment are calculated respectively using the dimensionless hull viscous force nonlinear model.

[0063] Specifically, the hull viscosity (including the longitudinal viscosity X H and the lateral viscous force Y H ), viscous moment N H The nonlinear expression of adopts the Inoue model, as shown below:

[0064]

[0065] Among them, X u ,Y v The equal terms are the fluid dynamic derivatives.

[0066] The forward speed, lateral speed and bow angular velocity are processed dimensionlessly, that is, the dimensionless forward speed is calculated based on the resultant speed and the forward speed; the dimensionless lateral speed is calculated based on the resultant speed and the lateral speed; and the dimensionless bow angular velocity is calculated based on the resultant speed, the ship length and the bow angular velocity; and the force and moment acting on the ship are processed dimensionlessly, as shown in formula (8):

[0067]

[0068] Dimensionless viscous force and moment of the hull (i.e. dimensionless longitudinal viscous force X' H , dimensionless transverse viscous force Y' H and the dimensionless bow viscous moment N' H The nonlinear model of ) is shown in formula (9):

[0069]

[0070] First, based on the ship's length, width and average draft, the fluid dynamics derivatives are calculated using empirical formulas. The fluid dynamics derivatives include linear fluid dynamics derivatives, second-order coupled derivatives, the first group of nonlinear fluid dynamics derivatives and the second group of nonlinear fluid dynamics derivatives. v , lateral force to bow angular velocity derivative Y' r , the lateral velocity derivative of the bow moment N' v , the derivative of the bow moment to the bow angular velocity N' r ), the Inoue formula is used to solve it, as shown in formula (10).

[0071]

[0072] Where λ = 2d m / L,d m is the average draft, d m =(d A +d F ) / 2,d F is the bow draft, d A is the stern draft, τ'=(d A -d F ) / d m is the dimensionless draft difference, the coefficient l v The expression is

[0073] Dimensionless longitudinal viscous force X' H As shown in formula (11):

[0074] X' H =X'(u')+X' vv v'2 +X' vr v'r'+X' rr r' 2 (11)

[0075] In the above formula, X'(u') is the straight-line resistance coefficient, S is the wetted area of ​​the hull, and C t is the total resistance coefficient of the ship.

[0076] Among them, according to the hull wetted area S, ship length L, average draft d m And the dimensionless forward speed u' is used to calculate the straight-line drag coefficient X'(u'), that is,

[0077] The hull wetted area S can be calculated using the Sandy formula (12):

[0078]

[0079] in, is the displacement volume, L W is the design waterline length, k is the area coefficient, and C is the mid-section coefficient. M Draft ratio to ship's breadth B / d m The function is obtained by querying the graph.

[0080] The total resistance coefficient of the ship is calculated as shown in formula (13):

[0081] C t =C f +C r +ΔC AR (13)

[0082] Among them, C f is the friction resistance coefficient, C r is the residual resistance coefficient, ΔC AR is the roughness compensation coefficient.

[0083] The friction resistance coefficient is calculated according to the 1957ITTC formula (14):

[0084]

[0085] Among them, R n is the Reynolds number of the ship.

[0086] The residual resistance coefficient can be calculated by the regression formula of the American series 60, and the roughness compensation coefficient ΔC AR Obtained by querying the following table.

[0087] Captain (m) <![CDATA[ΔC AR ]]> Captain (m) <![CDATA[ΔC AR ]]> 50~150 0.0004~0.00035 260~300 0 150~210 0.0002 300~350 -0.0001 210~260 0.0001 350~450 -0.00025

[0088] According to the lateral additional mass my Calculate the second-order coupling derivative X' vr , X' vr The calculation is shown in formula (15):

[0089] (X' vr +m y ) / m y =c m (15)

[0090] Among them, c m Calculated by Matsumoto's formula and Nagashima's formula (16):

[0091] c m (τ')=(1.11C b -0.07)(1+0.208τ') (16)

[0092] The first set of nonlinear hydrodynamic derivatives and the second set of nonlinear hydrodynamic derivatives can be calculated from Matsumoto's formula (17), that is, based on the ship's length, ship's width and average draft, the first set of nonlinear hydrodynamic derivatives (including the second-order lateral velocity derivative X') are calculated using Matsumoto's formula. vv and the second-order bow angular velocity derivative X' rr ), as shown below:

[0093]

[0094] Based on the ship length, ship width and average draft, the second set of nonlinear fluid dynamic derivatives (including the first high-order nonlinear derivative Y') is calculated using Zhou Zhaoming's regression formula. vv , the first mixed absolute value derivative Y' rr , the second mixed absolute value derivative Y' vr , the second higher-order nonlinear derivative N' vv , the first high-order mixed derivative N' rr and the second higher-order mixed derivative N' vr ), as shown in formula (18):

[0095]

[0096] 4. Calculation steps for forces and moments acting on the bare hull: Calculate the longitudinal force acting on the bare hull based on the longitudinal fluid inertia force and the dimensionless longitudinal viscous force; calculate the lateral force acting on the bare hull based on the lateral fluid inertia force and the dimensionless lateral viscous force; and calculate the rotational moment acting on the bare hull around the center of gravity of the ship based on the bow fluid inertia moment and the dimensionless bow viscous moment.

[0097] 5. Calculation steps for the force and torque acting on the propeller: Construct a linear correction model based on the thrust direction angle, basic derating coefficient and directional correction coefficient of the propeller, and use the linear correction model to correct the thrust derating coefficient of the propeller to obtain the corrected thrust derating coefficient; calculate the longitudinal thrust of the propeller and the lateral thrust of the propeller according to the total thrust of the propeller, the thrust direction angle of the propeller, the corrected thrust derating coefficient, the basic derating coefficient and the directional correction coefficient, and then obtain the longitudinal force acting on the propeller and the lateral force acting on the propeller; then calculate the rotational torque around the center of gravity of the ship acting on the propeller according to the position coordinates of the propeller in the hull coordinate system, the longitudinal thrust of the propeller and the lateral thrust of the propeller.

[0098] Specifically, the fluid dynamic model of the propeller is shown in formula (19):

[0099]

[0100] Among them, t P is the thrust derating coefficient, T is the propeller thrust, and the thrust derating coefficient is calculated according to the Shang He formula. The definition of propeller thrust is shown in formula (20).

[0101]

[0102] Among them, k T is the thrust coefficient, ρ is the fluid density, n is the propeller speed, D p is the diameter of the propeller.

[0103] The thrust coefficient is calculated according to the formula of Yoshimura and Nomoto (21):

[0104] k T =a0+a1J (21)

[0105] Among them, the coefficients a0 and a1 are obtained according to the propeller open water characteristic curve, J is the advance coefficient, J = V A / nD p , V A is the speed, V A =(1-w p )V, V is the ship speed, w p is the wake fraction at the propeller, which is calculated according to the Holtlow formula.

[0106] Based on the above-mentioned single-propeller hydrodynamic model, it is assumed that the arrangement distance of each propeller is far enough and there is no influence between them. Therefore, the following hydrodynamic model of a ship with multiple azimuth propellers is proposed, as shown in formula (22):

[0107]

[0108] Among them, X i represents the longitudinal thrust of the i-th propeller, Y i is the thrust generated by the i-th thruster in the lateral direction, N i represents the moment generated by the i-th thruster about the center of gravity of the ship, T i represents the total thrust generated by the i-th thruster, α i is the thrust direction angle of the propeller (0° points to the bow, rotating clockwise), (x i ,y i ) is the position of the propeller in the hull coordinate system, is the thrust derating coefficient of the i-th thruster (i.e. the corrected thrust derating coefficient).

[0109] The thrust reduction is caused by the change in pressure distribution around the hull due to the rotation of the propeller. The flow field distribution will be different when the propeller thrust direction angle is different, so the corresponding thrust reduction coefficient will change. Assuming that this change can be used with a linear model Indicates that it is based on the thrust direction angle of the propeller and the basic reduction coefficient b pi and the direction correction factor k pi Construct a linear correction model, called the correction term, and use the linear correction model (correction term) to correct the thrust reduction coefficient of the propeller to obtain the corrected thrust reduction coefficient, that is, the thrust reduction coefficient of the conventional propeller is used. Multiplication correction term Indicates the thrust derating factor after correction for the corresponding propeller.

[0110] 6. Calculation steps for the forces and moments acting on the rudder: Calculate the longitudinal force acting on the rudder based on the rudder's normal pressure, rudder angle, and rudder resistance reduction coefficient, and calculate the lateral force acting on the rudder based on the rudder's normal pressure and rudder angle. Calculate the rotational moment acting on the rudder about the ship's center of gravity based on the rudder's normal pressure, rudder angle, the distance from the point of action of the rudder's lateral force to the ship's center of gravity, and the distance from the point of action of the rudder's normal pressure to the ship's center of gravity.

[0111] Specifically, the fluid dynamics model on the rudder is shown in Equation (23):

[0112]

[0113] Among them, F N is the positive pressure of the rudder, δ is the rudder angle, t R The rudder resistance reduction coefficient is usually given by t R =0.29, a H In order to take into account the steering induction, the correction factor of the rudder force after the hull transverse force can be calculated according to formula (24), x His the distance from the point of action of the steering-induced hull transverse force (also called the action center of gravity) to the center of gravity of the ship, which can be calculated according to formula (25), x R It is the distance from the point of action of the rudder force (positive pressure of the rudder) to the center of gravity of the ship, which is approximately 0.5L.

[0114]

[0115] x H =-(0.4+0.1C b )L (25)

[0116] Rudder positive pressure F N The calculation is shown in formula (26):

[0117]

[0118] in, is the slope of the rudder lift coefficient when α = 0, A R is the rudder area, U R is the incoming flow velocity, α is the angle of attack, f α The Prandtl formula is used to calculate , as shown in formula (27):

[0119]

[0120] Among them, the incoming flow speed of the rudder u R is the longitudinal effective velocity of the incoming flow, v R is the effective lateral velocity of the incoming flow, and u is calculated using the Nomoto-Yoshimura method. R :

[0121]

[0122] Among them, u p is the effective flow velocity into the propeller, η=D p / H,D p is the propeller diameter, H is the height of the rudder, κ=0.6 / ε, ε=(1-w R0 ) / (1-w p0 ), w R0 is the wake coefficient of the rudder when the ship is sailing straight, w p0 is the wake coefficient of the propeller when the ship is sailing straight, s is the slip of the propeller, s=1-u p / nP, P is the pitch of the propeller.

[0123] v R It can be expressed as v R =v Rp -γ R (v+l R r), where vRp is the average lateral velocity caused by the asymmetric effect of the propeller's unidirectional rotation when the rudder is straight, (v+l R r) is the lateral velocity component of the water flow at the rudder caused by the ship's rotation and skew, l R ≈-0.9L~1.0L, γ R The calculation of is shown in formula (29):

[0124]

[0125] Attack angle α R The calculation is shown in formula (30), where β is the ship's drift angle, and the calculation is shown in formula (31):

[0126]

[0127] According to equations (23)-(31), the forces acting on the rudder can be calculated, that is, the lateral force, longitudinal force and rotational torque acting on the rudder and around the center of gravity of the ship can be calculated. Combined with the characteristics of the steering gear, the motion characteristics of the ship can be simulated. The characteristics of the steering gear are usually expressed using equation (32).

[0128]

[0129] Among them, δ E is the command rudder angle, T E is the servo time constant.

[0130] 7. Steps for calculating ship status data: Based on the longitudinal force, lateral force and rotational moment about the ship's center of gravity acting on the bare hull at the current moment, the longitudinal force, lateral force and rotational moment about the ship's center of gravity acting on the propeller, and the longitudinal force, lateral force and rotational moment about the ship's center of gravity acting on the rudder, and using the ship's motion equation to calculate the forward acceleration, lateral acceleration and bow angular acceleration at the current moment, then perform time integration operations on the forward acceleration, lateral acceleration and bow angular acceleration at the current moment to obtain the forward velocity, lateral velocity and bow angular velocity at the next moment after the current moment, and according to the forward velocity, lateral velocity and bow angular velocity at the next moment after the current moment, use the Runge-Kutta method to calculate the ship status data at the next moment after the current moment.

[0131] Specifically, if Figure 2 As shown, first, based on the longitudinal force, transverse force and rotational moment around the ship's center of gravity acting on the bare hull at the current moment (t), the longitudinal force, transverse force and rotational moment around the ship's center of gravity acting on the propeller, and the longitudinal force, transverse force and rotational moment around the ship's center of gravity acting on the rudder, the forward acceleration at the current moment (t) is calculated using the ship motion equation (also known as the ship maneuvering equation) lateral acceleration and bow angular acceleration Then, the forward acceleration, lateral acceleration and bow angular acceleration at the current moment are respectively integrated to obtain the forward speed, lateral speed and bow angular velocity at the next moment (t+1 moment) of the current moment. According to the forward speed, lateral speed and bow angular velocity at the next moment (t+1 moment) of the current moment, the Runge-Kutta method is used to calculate the ship state data at the next moment (t+1 moment) of the current moment. (Ship status data is the ship's position and heading), that is, the state prediction value at time t+1 is calculated.

[0132] 8. Steps for constructing the ship maneuvering motion model of an azimuth thruster: take the basic deduction coefficient, directional correction coefficient, straight-line resistance coefficient and fluid dynamic derivative as target optimization variables, construct the objective function based on the target optimization variables, the ship status data at the next moment of the current moment and the actual ship status data, and use the Adam optimization algorithm to calculate the optimal solution of the objective function, and then obtain the optimal target optimization variables; construct the ship maneuvering motion model of an azimuth thruster based on the optimal target optimization variables to simulate the maneuvering characteristics of the azimuth thruster and realize the training of ship drivers.

[0133] Specifically, the correlation coefficient in the manipulative model is first used as the target optimization variable, that is, the basic reduction coefficient b pi , Direction correction coefficient k pi , straight-line resistance coefficient and fluid dynamic derivative as the target optimization variables, and the target optimization variable is recorded as θ; then based on the target optimization variable θ, the next moment ship state data of the current moment And the real ship status data x t+1 Construct the objective function, that is, calculate the ship status data at the next moment (t+1 moment) of the current moment The error is calculated by comparing the actual ship status data at time t+1 obtained from the navigation data playback to find and the true value x t+1 The loss is minimized and the objective function is shown as follows:

[0134]

[0135] The Adam (Adaptive Moment Estimation) optimization algorithm is then used to calculate the optimal solution to the objective function. The parameters to be optimized are then updated by gradient backpropagation until the set number of gradient backpropagations is reached, resulting in the optimal target optimization variable. Finally, a maneuvering motion model for an azimuth thruster is constructed based on the optimal target optimization variable to simulate the maneuvering characteristics of the azimuth thruster and facilitate ship crew training. The Adam optimization algorithm adaptively adjusts the learning rates of each parameter in the target optimization variable using the first-order moment (gradient mean) and second-order moment (gradient mean), eliminating the complexity of manual parameter tuning and addressing multi-parameter dimensional differences and gradient heterogeneity. Adjustment is performed based on the first-order and second-order moment estimates of the gradient, accelerating parameter convergence. Furthermore, the Adam optimization algorithm performs an exponentially weighted moving average of the first-order and second-order moment estimates of the gradient, incorporating historical gradient information into the parameter update. This facilitates smooth parameter updates and is robust to hyperparameter selection, achieving good performance without requiring extensive parameter tuning.

[0136] The present invention also relates to an azimuth propeller simulation control system based on empirical formulas and based on empirical formula priors. The system corresponds to the above-mentioned azimuth propeller simulation control method based on empirical formulas and based on empirical formula priors, and can be understood as a system for implementing the above-mentioned method. The system includes a parameter acquisition module, a fluid inertia force and torque calculation module, a fluid viscosity force and torque calculation module, a force and torque calculation module acting on the bare hull, a force and torque calculation module acting on the propeller, a force and torque calculation module acting on the rudder, a ship status data calculation module, and an azimuth propeller ship maneuvering motion model construction module, which are connected in sequence. Specifically,

[0137] The parameter acquisition module acquires ship operation parameters, propeller parameters and rudder parameters, and establishes a hull coordinate system with the ship's center of gravity as the origin, the direction of the ship's bow as the longitudinal axis, the direction of the ship's starboard side as the transverse axis, and the vertically downward direction of the ship's bottom as the vertical axis; the ship operation parameters include ship mass, ship length, ship width, bow draft, stern draft, hull wetted area, ship's forward speed, lateral speed and bow angular velocity; the actual motion resultant speed of the ship is calculated based on the forward speed and lateral speed, and the average draft is calculated based on the bow draft and stern draft; the propeller parameters include the total thrust of the propeller, the thrust direction angle, the basic deduction coefficient, the direction correction coefficient, the propeller thrust deduction coefficient and the position coordinates of the propeller in the hull coordinate system; the rudder parameters include the rudder's positive pressure, rudder angle, rudder resistance deduction coefficient, the distance from the point of action of the rudder's lateral force to the ship's center of gravity, and the longitudinal distance between the point of action of the rudder force and the ship's center of gravity;

[0138] The fluid inertia force and moment calculation module calculates the longitudinal added mass and the transverse added mass based on the ship mass, ship length, ship width and average draft, and adopts the ship fluid dynamic derivative regression formula in the empirical formula; calculates the vertical additional moment of inertia based on the ship length and ship width, and adopts the ship fluid dynamic derivative regression formula; calculates the longitudinal fluid inertia force and the transverse fluid inertia force based on the longitudinal added mass, the transverse added mass, the forward speed, the lateral speed and the bow angular velocity, and calculates the bow fluid inertia moment based on the vertical additional moment of inertia, the longitudinal added mass, the transverse added mass, the forward speed, the lateral speed and the bow angular velocity;

[0139] The fluid viscosity force and torque calculation module calculates the dimensionless forward velocity according to the resultant velocity and the forward velocity; calculates the dimensionless lateral velocity according to the resultant velocity and the lateral velocity; and calculates the dimensionless bow angular velocity according to the resultant velocity, the ship length, and the bow angular velocity; calculates the straight-line resistance coefficient according to the hull wetted area, the ship length, the average draft, and the dimensionless forward velocity, and then calculates the fluid dynamic derivative based on the ship length, the ship width, and the average draft using an empirical formula; and calculates the dimensionless longitudinal viscous force, the dimensionless transverse viscous force, and the dimensionless bow viscous moment based on the straight-line resistance coefficient, the fluid dynamic derivative, the dimensionless forward velocity, the dimensionless lateral velocity, and the dimensionless bow angular velocity using a dimensionless hull viscosity nonlinear model;

[0140] The force and moment calculation module acting on the bare hull calculates the longitudinal force acting on the bare hull based on the longitudinal fluid inertia force and the dimensionless longitudinal viscous force; calculates the lateral force acting on the bare hull based on the lateral fluid inertia force and the dimensionless lateral viscous force; and calculates the rotational moment acting on the bare hull around the center of gravity of the ship based on the bow fluid inertia moment and the dimensionless bow viscous moment;

[0141] The force and torque calculation module acting on the propeller constructs a linear correction model based on the thrust direction angle, basic deduction coefficient and direction correction coefficient of the propeller, and uses the linear correction model to correct the thrust deduction coefficient of the propeller to obtain a corrected thrust deduction coefficient; according to the total thrust of the propeller, the thrust direction angle of the propeller, the corrected thrust deduction coefficient, the basic deduction coefficient and the direction correction coefficient, the longitudinal thrust of the propeller and the transverse thrust of the propeller are calculated respectively, thereby obtaining the longitudinal force acting on the propeller and the transverse force acting on the propeller; and then, according to the position coordinates of the propeller in the hull coordinate system, the longitudinal thrust of the propeller and the transverse thrust of the propeller, the rotational torque around the center of gravity of the ship is calculated;

[0142] The rudder force and torque calculation module calculates the longitudinal force acting on the rudder based on the rudder's normal pressure, rudder angle, and rudder resistance deduction coefficient, calculates the lateral force acting on the rudder based on the rudder's normal pressure and rudder angle, and calculates the rotational torque acting on the rudder about the ship's center of gravity based on the rudder's normal pressure, rudder angle, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's normal pressure application point to the ship's center of gravity;

[0143] The ship state data calculation module calculates the forward acceleration, lateral acceleration and bow angular acceleration at the current moment based on the longitudinal force, lateral force and rotational torque about the ship's center of gravity acting on the bare hull, the longitudinal force, lateral force and rotational torque about the ship's center of gravity acting on the propeller, and the longitudinal force, lateral force and rotational torque about the ship's center of gravity acting on the rudder at the current moment, and uses the ship motion equation to calculate; then, time integration operations are performed on the forward acceleration, lateral acceleration and bow angular acceleration at the current moment to obtain the forward velocity, lateral velocity and bow angular velocity at the next moment of the current moment; based on the forward velocity, lateral velocity and bow angular velocity at the next moment of the current moment, the Runge-Kutta method is used to calculate the ship state data at the next moment of the current moment;

[0144] The module for constructing a ship maneuvering motion model for an azimuth thruster uses a basic deduction coefficient, a directional correction coefficient, a straight-line resistance coefficient, and a fluid dynamic derivative as target optimization variables, constructs a target function based on the target optimization variables, the ship state data at the next moment of the current moment, and the actual ship state data, and uses the Adam optimization algorithm to calculate the optimal solution of the target function, thereby obtaining the optimal target optimization variables; the module for constructing a ship maneuvering motion model for an azimuth thruster is constructed based on the optimal target optimization variables to simulate the maneuvering characteristics of the azimuth thruster and realize the training of ship drivers.

[0145] Preferably, in the fluid viscous force and torque calculation module, the fluid dynamic derivative is calculated based on the ship length, ship width and average draft using an empirical formula, specifically including:

[0146] The empirical formulas include the Inoue formula, the Matsumoto formula and the ship fluid dynamic derivative regression formula. The linear fluid dynamic derivative is calculated based on the ship length, ship width and average draft using the Inoue formula; the second-order coupling derivative is calculated based on the lateral added mass; the first group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the Matsumoto formula; the second group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the ship fluid dynamic derivative regression formula; and the fluid dynamic derivative is generated based on the linear fluid dynamic derivative, the second-order coupling derivative, the first group of nonlinear fluid dynamic derivatives and the second group of nonlinear fluid dynamic derivatives.

[0147] Preferably, the linear fluid dynamic derivatives include the derivative of lateral force to lateral velocity, the derivative of lateral force to bow angular velocity, the derivative of bow moment to lateral velocity, and the derivative of bow moment to bow angular velocity; the first group of nonlinear fluid dynamic derivatives include second-order lateral velocity derivatives and second-order bow angular velocity derivatives, and the second group of nonlinear fluid dynamic derivatives include first high-order nonlinear derivatives, second high-order nonlinear derivatives, first mixed absolute value derivatives, second mixed absolute value derivatives, first high-order mixed derivatives, and second high-order mixed derivatives.

[0148] Preferably, the ship status data includes the ship's position and heading.

[0149] Preferably, the Adam optimization algorithm adaptively adjusts the learning rate of each parameter in the target optimization variable through the first-order moment and the second-order moment to avoid the complexity of manual parameter adjustment.

[0150] The present invention provides an objective and scientific azimuth propeller simulation control method and system based on empirical formulas and prior knowledge. By correcting the propeller thrust reduction coefficient, a fluid dynamic model of the azimuth propeller is established, which more accurately simulates the maneuvering motion of a ship equipped with an azimuth propeller, providing a scene closer to actual ship maneuvering for crew training, and facilitating large-scale teaching and demonstration. By combining empirical formulas to obtain a rough estimate of the ship's maneuverability coefficient, introducing prior knowledge, and then using the Adam optimization algorithm and combining a small amount of ship maneuvering data to further fit the maneuverability coefficient, a realistic ship motion simulation is achieved by comprehensively considering multiple factors such as fluid inertia force, viscous force, propeller thrust, and rudder force, greatly reducing the experimental cost of fitting the maneuverability coefficient and the cost of building the simulation system.

[0151] It should be noted that the specific embodiments described above can enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although this specification has described the present invention in detail with reference to the drawings and embodiments, those skilled in the art should understand that the present invention can still be modified or replaced with equivalents. In short, all technical solutions and improvements that do not depart from the spirit and scope of the present invention should be included in the scope of protection of the patent for the present invention.

Claims

1. A simulation control method for an azimuth thruster based on an empirical formula prior, characterized in that: The following steps are involved: Parameter acquisition step: acquiring ship operation parameters, propeller parameters and rudder parameters, and establishing a hull coordinate system with the ship's center of gravity as the origin, the direction of the ship's bow as the longitudinal axis, the direction of the ship's starboard side as the transverse axis, and the vertically downward direction of the ship's bottom as the vertical axis; the ship operation parameters include ship mass, ship length, ship width, bow draft, stern draft, hull wetted area, ship's forward speed, lateral speed and bow angular velocity; calculating the combined speed of the ship's actual motion based on the forward speed and lateral speed, and calculating the average draft based on the bow draft and stern draft; the propeller parameters include the total thrust of the propeller, the thrust direction angle, the basic deduction coefficient, the direction correction coefficient, the propeller's thrust deduction coefficient and the position coordinates of the propeller in the hull coordinate system; the rudder parameters include the rudder's positive pressure, rudder angle, rudder resistance deduction coefficient, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's positive pressure application point to the ship's center of gravity; Calculation steps for fluid inertia force and moment: Based on the ship's mass, length, width and average draft, the longitudinal added mass and transverse added mass are calculated using empirical formulas; based on the ship's length and width, the vertical added moment of inertia is calculated using empirical formulas; based on the longitudinal added mass, transverse added mass, forward speed, lateral speed and bow angular velocity, the longitudinal fluid inertia force and transverse fluid inertia force are calculated respectively, and the bow fluid inertia moment is calculated based on the vertical added moment of inertia, the longitudinal added mass, transverse added mass, forward speed, lateral speed and bow angular velocity; Calculation steps for fluid viscous force and torque: Calculate the dimensionless forward velocity based on the resultant velocity and forward velocity; calculate the dimensionless lateral velocity based on the resultant velocity and lateral velocity; and calculate the dimensionless bow angular velocity based on the resultant velocity, ship length, and bow angular velocity; calculate the straight-line resistance coefficient based on the hull wetted area, ship length, average draft, and dimensionless forward velocity, and then calculate the fluid dynamic derivative based on the ship length, ship width, and average draft using an empirical formula; based on the straight-line resistance coefficient, fluid dynamic derivative, dimensionless forward velocity, dimensionless lateral velocity, and dimensionless bow angular velocity, the dimensionless longitudinal viscous force, dimensionless transverse viscous force, and dimensionless bow viscous moment are calculated using the dimensionless hull viscous force nonlinear model; Calculation steps for the forces and moments acting on the bare hull: Calculate the longitudinal force acting on the bare hull based on the longitudinal fluid inertia force and the dimensionless longitudinal viscous force; calculate the transverse force acting on the bare hull based on the transverse fluid inertia force and the dimensionless transverse viscous force; and calculate the rotational moment acting on the bare hull about the center of gravity of the ship based on the bow fluid inertia moment and the dimensionless bow viscous moment; Calculation steps for the force and torque acting on the propeller: construct a linear correction model based on the thrust direction angle, basic derating coefficient and directional correction coefficient of the propeller, and use the linear correction model to correct the thrust derating coefficient of the propeller to obtain the corrected thrust derating coefficient; calculate the longitudinal thrust of the propeller and the transverse thrust of the propeller according to the total thrust of the propeller, the thrust direction angle of the propeller, the corrected thrust derating coefficient, the basic derating coefficient and the directional correction coefficient, and then obtain the longitudinal force acting on the propeller and the transverse force acting on the propeller; then calculate the rotational torque acting on the propeller around the center of gravity of the ship according to the position coordinates of the propeller in the hull coordinate system, the longitudinal thrust of the propeller and the transverse thrust of the propeller; Calculation steps for the forces and moments acting on the rudder: Calculate the longitudinal force acting on the rudder based on the rudder's normal pressure, rudder angle, and rudder resistance reduction coefficient. Calculate the lateral force acting on the rudder based on the rudder's normal pressure and rudder angle. Calculate the rotational moment acting on the rudder about the ship's center of gravity based on the rudder's normal pressure, rudder angle, the distance from the rudder's point of application of the lateral force to the ship's center of gravity, and the distance from the rudder's point of application of the normal pressure to the ship's center of gravity. Ship status data calculation steps: Based on the longitudinal force, transverse force and rotational moment about the ship's center of gravity acting on the bare hull at the current moment, the longitudinal force, transverse force and rotational moment about the ship's center of gravity acting on the propeller, and the longitudinal force, transverse force and rotational moment about the ship's center of gravity acting on the rudder, and using the ship's motion equation, calculate the forward acceleration, lateral acceleration and bow angular acceleration at the current moment; Then, the forward acceleration, lateral acceleration and bow angular acceleration at the current moment are respectively integrated over time to obtain the forward speed, lateral speed and bow angular velocity at the next moment from the current moment. Based on the forward speed, lateral speed and bow angular velocity at the next moment from the current moment, the Runge-Kutta method is used to calculate the ship state data at the next moment from the current moment. The steps for constructing the maneuvering motion model of an azimuth thruster ship are as follows: the basic deduction coefficient, directional correction coefficient, straight-line resistance coefficient, and hydrodynamic derivative are used as target optimization variables. The objective function is constructed based on the target optimization variables, the ship state data at the next moment of the current moment, and the actual ship state data. The Adam optimization algorithm is used to calculate the optimal solution of the objective function, thereby obtaining the optimal target optimization variables. A ship maneuvering motion model of an azimuth thruster is constructed based on the optimal objective optimization variables to simulate the maneuvering characteristics of the azimuth thruster.

2. The method for simulating and controlling an azimuth thruster based on empirical formulas according to claim 1, characterized in that: In the fluid viscous force and moment calculation step, the fluid dynamic derivative is calculated based on the ship length, ship width and average draft using an empirical formula, specifically including: The empirical formulas include the Inoue formula, the Matsumoto formula and the ship fluid dynamic derivative regression formula. The linear fluid dynamic derivative is calculated based on the ship length, ship width and average draft using the Inoue formula; the second-order coupling derivative is calculated based on the lateral added mass; the first group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the Matsumoto formula; the second group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the ship fluid dynamic derivative regression formula; and the fluid dynamic derivative is generated based on the linear fluid dynamic derivative, the second-order coupling derivative, the first group of nonlinear fluid dynamic derivatives and the second group of nonlinear fluid dynamic derivatives.

3. The method for simulating and controlling an azimuth thruster based on empirical formulas according to claim 2, characterized in that: The linear fluid dynamic derivatives include the derivative of lateral force to lateral velocity, the derivative of lateral force to bow angular velocity, the derivative of bow moment to lateral velocity, and the derivative of bow moment to bow angular velocity; the first group of nonlinear fluid dynamic derivatives include second-order lateral velocity derivatives and second-order bow angular velocity derivatives, and the second group of nonlinear fluid dynamic derivatives include first high-order nonlinear derivatives, second high-order nonlinear derivatives, first mixed absolute value derivatives, second mixed absolute value derivatives, first high-order mixed derivatives, and second high-order mixed derivatives.

4. The method for simulating and controlling an azimuth thruster based on empirical formulas according to claim 1, characterized in that: In the ship status data calculation step, the ship status data includes the ship's position and heading.

5. The method for simulating and controlling an azimuth thruster based on empirical formulas according to claim 1, characterized in that: In the step of constructing the full-turn propeller ship maneuvering motion model, the Adam optimization algorithm adaptively adjusts the learning rate of each parameter in the target optimization variable through the first-order moment and the second-order moment to avoid the complexity of manual parameter adjustment.

6. A simulation control system for an azimuth thruster based on an empirical formula prior, characterized in that: It includes a parameter acquisition module, a fluid inertia force and torque calculation module, a fluid viscous force and torque calculation module, a force and torque calculation module acting on the bare hull, a force and torque calculation module acting on the propeller, a force and torque calculation module acting on the rudder, a ship status data calculation module and an azimuth thruster ship maneuvering motion model construction module, which are connected in sequence. The parameter acquisition module acquires ship operation parameters, propeller parameters and rudder parameters, and establishes a hull coordinate system with the ship's center of gravity as the origin, the direction of the ship's bow as the longitudinal axis, the direction of the ship's starboard side as the transverse axis, and the vertically downward direction of the ship's bottom as the vertical axis; the ship operation parameters include ship mass, ship length, ship width, bow draft, stern draft, hull wetted area, ship's forward speed, lateral speed and bow angular velocity; the actual motion resultant speed of the ship is calculated based on the forward speed and lateral speed, and the average draft is calculated based on the bow draft and stern draft; the propeller parameters include the total thrust of the propeller, the thrust direction angle, the basic deduction coefficient, the direction correction coefficient, the propeller thrust deduction coefficient and the position coordinates of the propeller in the hull coordinate system; the rudder parameters include the rudder's positive pressure, rudder angle, rudder resistance deduction coefficient, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's positive pressure application point to the ship's center of gravity; The fluid inertia force and moment calculation module calculates the longitudinal added mass and the transverse added mass based on the ship mass, ship length, ship width and average draft using empirical formulas; calculates the vertical added moment of inertia based on the ship length and ship width using empirical formulas; calculates the longitudinal fluid inertia force and the transverse fluid inertia force based on the longitudinal added mass, the transverse added mass, the forward speed, the lateral speed and the bow angular velocity, and calculates the bow fluid inertia moment based on the vertical added moment of inertia, the longitudinal added mass, the transverse added mass, the forward speed, the lateral speed and the bow angular velocity; The fluid viscosity force and torque calculation module calculates the dimensionless forward velocity according to the resultant velocity and the forward velocity; calculates the dimensionless lateral velocity according to the resultant velocity and the lateral velocity; and calculates the dimensionless bow angular velocity according to the resultant velocity, the ship length, and the bow angular velocity; calculates the straight-line resistance coefficient according to the hull wetted area, the ship length, the average draft, and the dimensionless forward velocity, and then calculates the fluid dynamic derivative based on the ship length, the ship width, and the average draft using an empirical formula; and calculates the dimensionless longitudinal viscous force, the dimensionless transverse viscous force, and the dimensionless bow viscous moment based on the straight-line resistance coefficient, the fluid dynamic derivative, the dimensionless forward velocity, the dimensionless lateral velocity, and the dimensionless bow angular velocity using a dimensionless hull viscosity nonlinear model; The force and moment calculation module acting on the bare hull calculates the longitudinal force acting on the bare hull based on the longitudinal fluid inertia force and the dimensionless longitudinal viscous force; The lateral force acting on the bare hull is calculated based on the lateral fluid inertia force and the dimensionless lateral viscous force; and the rotational moment acting on the bare hull around the center of gravity of the ship is calculated based on the bow fluid inertia moment and the dimensionless bow viscous moment; The force and torque calculation module acting on the propeller constructs a linear correction model based on the thrust direction angle, basic deduction coefficient and direction correction coefficient of the propeller, and uses the linear correction model to correct the thrust deduction coefficient of the propeller to obtain a corrected thrust deduction coefficient; according to the total thrust of the propeller, the thrust direction angle of the propeller, the corrected thrust deduction coefficient, the basic deduction coefficient and the direction correction coefficient, the longitudinal thrust of the propeller and the transverse thrust of the propeller are calculated respectively, thereby obtaining the longitudinal force acting on the propeller and the transverse force acting on the propeller; and then, according to the position coordinates of the propeller in the hull coordinate system, the longitudinal thrust of the propeller and the transverse thrust of the propeller, the rotational torque around the center of gravity of the ship is calculated; The rudder force and torque calculation module calculates the longitudinal force acting on the rudder based on the rudder's normal pressure, rudder angle, and rudder resistance deduction coefficient, calculates the lateral force acting on the rudder based on the rudder's normal pressure and rudder angle, and calculates the rotational torque acting on the rudder about the ship's center of gravity based on the rudder's normal pressure, rudder angle, the distance from the rudder's lateral force application point to the ship's center of gravity, and the distance from the rudder's normal pressure application point to the ship's center of gravity; The ship status data calculation module calculates the forward acceleration, lateral acceleration, and bow angular acceleration at the current moment based on the longitudinal force, lateral force, and rotational moment about the ship's center of gravity acting on the bare hull, the longitudinal force, lateral force, and rotational moment about the ship's center of gravity acting on the propeller, and the longitudinal force, lateral force, and rotational moment about the ship's center of gravity acting on the rudder, and uses the ship's motion equations; Then, the forward acceleration, lateral acceleration and bow angular acceleration at the current moment are respectively integrated over time to obtain the forward speed, lateral speed and bow angular velocity at the next moment from the current moment. Based on the forward speed, lateral speed and bow angular velocity at the next moment from the current moment, the Runge-Kutta method is used to calculate the ship state data at the next moment from the current moment. The module for constructing a ship maneuvering motion model for an azimuth thruster uses the basic deduction coefficient, directional correction coefficient, straight-line resistance coefficient, and hydrodynamic derivative as target optimization variables. An objective function is constructed based on the target optimization variables, the ship's state data for the next moment at the current moment, and the actual ship's state data. The Adam optimization algorithm is used to calculate the optimal solution of the objective function, thereby obtaining the optimal target optimization variables. A ship maneuvering motion model of an azimuth thruster is constructed based on the optimal objective optimization variables to simulate the maneuvering characteristics of the azimuth thruster.

7. The azimuth propeller simulation control system based on empirical formula priori according to claim 6, characterized in that: In the fluid viscous force and torque calculation module, the fluid dynamic derivatives are calculated based on the ship length, ship width and average draft using empirical formulas, specifically including: The empirical formulas include the Inoue formula, the Matsumoto formula and the ship fluid dynamic derivative regression formula. The linear fluid dynamic derivative is calculated based on the ship length, ship width and average draft using the Inoue formula; the second-order coupling derivative is calculated based on the lateral added mass; the first group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the Matsumoto formula; the second group of nonlinear fluid dynamic derivatives is calculated based on the ship length, ship width and average draft using the ship fluid dynamic derivative regression formula; and the fluid dynamic derivative is generated based on the linear fluid dynamic derivative, the second-order coupling derivative, the first group of nonlinear fluid dynamic derivatives and the second group of nonlinear fluid dynamic derivatives.

8. The azimuth propeller simulation control system based on empirical formula priori according to claim 7, characterized in that: The linear fluid dynamic derivatives include the derivative of lateral force to lateral velocity, the derivative of lateral force to bow angular velocity, the derivative of bow moment to lateral velocity, and the derivative of bow moment to bow angular velocity; the first group of nonlinear fluid dynamic derivatives include second-order lateral velocity derivatives and second-order bow angular velocity derivatives, and the second group of nonlinear fluid dynamic derivatives include first high-order nonlinear derivatives, second high-order nonlinear derivatives, first mixed absolute value derivatives, second mixed absolute value derivatives, first high-order mixed derivatives, and second high-order mixed derivatives.

9. The azimuth propeller simulation control system based on empirical formula priori according to claim 6, characterized in that: The ship status data includes the ship's position and heading.

10. The azimuth propeller simulation control system based on empirical formula priori according to claim 6, characterized in that: The Adam optimization algorithm adaptively adjusts the learning rate of each parameter in the target optimization variable through the first-order moment and the second-order moment to avoid the complexity of manual parameter adjustment.