A method and system for adjusting ballast water of a full-circle crane piling ship

Abstract

The application discloses a kind of full slewing crane pile driver ship ballast water regulating method and system. Including: establishing ship appendage coordinate system;Synchronous acquisition hoisting load, ship floating state and ballast tank state data;Build the full slewing dynamic moment balance mathematical model including longitudinal and lateral coupling effect and free surface correction;Ballast water regulation is converted into the strong constraint double-objective optimization model with the shortest deployment time and the largest stability reserve as target;Improved NSGA-II algorithm is used for iterative solution, including heuristic moment compensation initialization, safety margin weighted crowding distance calculation and feasibility priority tournament selection;Pareto optimal solution set is converted into ballast pump start-stop and valve opening degree instruction, and based on the deviation of real-time monitoring inclination and model prediction inclination triggers dynamic compensation.The application realizes the fast, accurate and safe generation of ballast water scheme, so as to realize the stability regulation of full slewing crane pile driver ship in complex dynamic construction environment with high response, high precision and intrinsic safety.

Description

Technical Field

[0001] This invention relates to the field of shipbuilding and marine engineering technology, and more specifically, to a method and system for adjusting ballast water on a fully rotating crane piling vessel. Background Technology

[0002] When a fully azimuth-rotor piling vessel performs offshore piling operations, the piling frame and the suspended piles need to undergo large-scale amplitude changes and full-rotation movements with the superstructure. This causes a violent and continuous dynamic drift in the ship's center of gravity, resulting in a huge variable capsizing moment. If the ballast water cannot be adjusted in real time and accurately to generate a reverse restoring moment, the ship will face the risk of loss of stability or even capsizing. Traditional ship ballast water regulation relies on manual experience, which has a significant lag. Existing intelligent regulation research mainly focuses on two directions: one is to establish a dynamic programming model with the goal of minimizing the allocation time, which involves mutual allocation within the ballast tanks while exchanging water with seawater; the other is to establish a mathematical model with the goal of minimizing the amount of water allocated, which involves mutual allocation of water within the ballast tanks while keeping the total amount allocated constant.

[0003] However, both of the above-mentioned adjustment methods have significant drawbacks in the application of slewing crane piling vessels. First, from the perspective of the response mechanism, both inter-compartment transfer and complex mixing involve complex pipeline topology switching. The flow rate is limited by the liquid level pressure difference between connected compartments, which means that the adjustment rate often cannot match the rapid slewing speed of the crane piling vessel during dynamic construction, contradicting the urgent need for rapid leveling at the construction site. Second, from the perspective of algorithm optimization, conventional algorithms are insufficient when dealing with multi-objective optimization problems such as those involving high-dimensional decision variables (multiple ballast tanks) and strong physical constraints (strict stability requirements) in crane piling vessels. For example, the simulated annealing (SA) algorithm, due to its single-point search mechanism, struggles to capture the Pareto front of "time-safety" and its slow cooling process fails to meet real-time requirements. The decomposition-based multi-objective evolutionary algorithm (MOEA / D), when faced with complex constraints, has a pre-defined weight vector that is difficult to adapt to the feasible solution manifold, easily leading to uneven distribution of the solution set or getting trapped in local optima. The standard non-dominated sorting genetic algorithm (NSGA-II), due to its random initialization strategy, generates more than 90% infeasible solutions under strong stability constraints, resulting in a huge waste of computing power. Furthermore, its selection mechanism based on crowding distance does not distinguish between objective priorities, making it difficult to lock in a highly stable solution within a limited time.

[0004] Currently, the contradiction between the high real-time requirements of attitude control for slewing crane pile-driving vessels and the low convergence efficiency of existing optimization algorithms, as well as the lack of an inherently safe proactive decision-making mechanism under strong physical constraints, severely restricts the improvement of the refined management level and the inherent safety efficiency of offshore pile-driving operations. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for adjusting ballast water on a fully rotating hoisting piling vessel, in order to solve the above-mentioned problems existing in the prior art.

[0006] The application is as follows: A method for regulating ballast water on a fully slewing hoisting piling vessel includes: Establish the ship's appendage coordinate system O-XYZ, with the intersection of the ship's baseline and the mid-longitudinal section as the X-axis and positive in the direction of the bow, the mid-transverse section as the Y-axis and positive in the direction of the port side, and the vertical axis upward as the Z-axis. Real-time data is collected synchronously at a preset period T, including hoisting load data, ship floating status data, and ballast tank status data. Based on the principles of ship statics, a set of nonlinear equilibrium equations, including the longitudinal and transverse heel coupling effect and free surface correction, is constructed to generate a full-rotation dynamic moment balance mathematical model. The set of nonlinear equilibrium equations is used to calculate the ballast moment required to maintain the ship's balance based on the lifting load data, the ship's floating state data, and the ballast tank state data, combined with the coordinate parameters in the ship's appendage coordinate system O-XYZ. Based on the ballast moment, the ballast water regulation is transformed into a bi-objective optimization problem with strong physical constraints. A multi-objective optimization model is constructed, which includes a first objective function of minimizing the allocation time and a second objective function of maximizing the stability reserve. Hard constraints on the safety tilt angle, physical capacity constraints, and unidirectional exchange logic constraints are set. The calculation of the first objective function and the second objective function depends on the parameters in the nonlinear equilibrium equation system. An improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model and output the Pareto optimal solution set. The initialization of the improved NSGA-II algorithm depends on the established nonlinear equilibrium equation system, and its individual evaluation depends on the constructed first objective function, second objective function, and various constraints. The output Pareto optimal solution set is converted into ballast pump start / stop and valve opening commands, and the improved NSGA-II algorithm is triggered to perform dynamic compensation based on the deviation between the real-time monitored tilt angle and the model predicted tilt angle in the collected real-time data.

[0007] Furthermore, the lifting load data includes the mass of the lifted object Pt, the current lifting radius R, and the real-time slewing angle α of the crane relative to the bow of the ship; the ship's buoyancy data includes the ship's current average draft T. m Instantaneous tilt angle θ real and pitch angle φ real The ballast tank status data includes the current water volume set V0 = [v...] of all n ballast tanks on the ship.{1,0} , v {2,0} ,..., v {n,0} The current water volume set V0 is used to calculate the ship's current total displacement.

[0008] Furthermore, the nonlinear equilibrium equation set specifically includes: Lateral moment balance equation:

[0009] Longitudinal moment balance equation:

[0010] in, This refers to the ship's current total displacement. ;Δ light This indicates the empty ship's displacement, and the current water volume v i,0 This data originates from the ballast tank status data; , These are the horizontal and vertical centering heights, respectively. The free surface correction value is given, KG is the height of the ship's center of gravity, and θ is the free surface correction value. These are the rake angle and the pitch angle, respectively. , The moment of inertia is the surface area of ​​the liquid inside the chamber; This refers to the water volume adjustment amount for the i-th ballast tank; , Let ρ be the volume center coordinate of the i-th ballast tank in the ship's appendage coordinate system O-XYZ, where ρ is the density of seawater and g is the acceleration due to gravity.

[0011] Furthermore, the multi-objective optimization model includes: The primary objective function is to minimize the deployment time and the maximum single-pump operating time.

[0012] Where Ω represents the set of ballast tanks involved in the regulation, i.e., Δv i ≠0 cabins; Q i η represents the actual flow rate of the pump unit corresponding to the i-th ballast tank. sys t is the volumetric efficiency coefficient of the pump set. value This refers to the valve opening and closing response delay time. The second objective function is to maximize the stability reserve, minimize the residual tilt angle, and suppress the free surface:

[0013] Where, θ obj , obj All are ideal target tilt angles; ω θ, All are weighting coefficients; S i Let λ be the free surface area inside the i-th ballast tank; λ be the penalty factor; and X be the decision variable vector. Constraint settings: Set hard constraints for safety tilt angle, physical capacity constraints, and unidirectional switching logic constraints; The safety tilt angle hard constraints include: tilt angle θ ≤ 3.5°, and pitch angle... If the value is ≤2°, the solution that violates this constraint is an infeasible solution; The physical capacity constraints include: 0≤v i,0 +Δv i ≤V max,i Wherein v i,0 V is derived from the ballast tank status data. max,i The maximum capacity of the i-th ballast tank; The unidirectional exchange logic constraints include: Δv i ·Δv j ≤0 is used to prioritize the direct exchange of ballast water with the outside, without inter-cabin water exchange. This represents the water volume adjustment amount for the j-th ballast tank.

[0014] Furthermore, the improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model, outputting a Pareto optimal solution set, including: S51. Heuristic Torque Compensation Initialization: Set the evolutionary generation counter t=0, set the population size to N, and the maximum number of iterations to G. max Based on the nonlinear equilibrium equations, the mass P of the lifted object, the current lifting radius R, and the real-time slewing angle α, the theoretical regulating water volume benchmark value of each ballast tank is derived in reverse, and the initial parent population P0 is generated based on the theoretical regulating water volume benchmark value. S52. Non-dominated sorting and safety margin weighted crowding distance calculation: For each individual in the population, evaluate it according to the first objective function and the second objective function, calculate its allocation time objective function value and stability reserve objective function value, and then execute the stratification mechanism; S53. Feasibility-first tournament selection: Select outstanding individuals from the parent population to enter the pairing pool, define the constraint violation degree CV(p) of individual p according to the hard constraint of the safety tilt angle, and execute the feasibility comparison rule. S54. Adaptive Genetic Operators and Physical Boundary Checks: Perform crossover and mutation operations on individuals entering the pairing pool to produce the offspring population Q. t And perform boundary correction on the new individual based on the physical capacity constraints; S55. Population Merging and Elite Preservation Strategy: Merge the current parent population P... t With the generated offspring population Q t Merge them to form a combined population R t For all individuals in the combined population, the non-dominated ranking and safety margin weighted crowding distance calculation described in S52 are re-executed, and individuals are selected sequentially in ascending order of non-dominated level to fill the new population P. t+1 ; S56. Iteration Termination Judgment and Result Output: Check whether the current generation t has reached the maximum number of iterations G. max If the target is not reached, set t=t+1 and return to S53 to continue iterating. If the target is reached, output the individual with the first level of non-dominance as the solution set of the Pareto optimal ballast allocation scheme.

[0015] Furthermore, the heuristic torque compensation initialization includes: Calculate the theoretical compensation water volume: Using the lateral moment balance equation and the longitudinal moment balance equation, and based on the lifted object mass P, the current lifting radius R, and the real-time slewing angle α, deduce in reverse the overturning moment M generated by the current slewing. req Required theoretical regulating water volume benchmark values ​​for each ballast tank :

[0016] Where β is the proportional gain coefficient, y i Let y be the Y-axis coordinate of the i-th ballast tank in the ship's appendage coordinate system O-XYZ; j Let be the Y-axis coordinate of the j-th ballast tank in the ship's appendage coordinate system O-XYZ; Generate seed individuals: Centered on the theoretical water regulation baseline value, a random perturbation term ε following a normal distribution N(0,σ²) is superimposed to generate heuristic seed individuals, which account for 30%~50% of the initial population; Population construction: The remaining individuals are randomly generated within the range that satisfies the set physical capacity constraints, and together with the heuristic seed individuals, they form the initial parent population P0.

[0017] Furthermore, the calculation of the non-dominated sorting and safety margin weighted crowding distance specifically includes: Fast non-dominated sorting: Based on the dominance relationship between individuals, the population is divided into different non-dominated levels, where the individuals in the first non-dominated level are the current optimal solution set; Calculating the weighted crowding distance with a safety margin: Within the same non-dominated hierarchy, by introducing a safety weighting factor and assigning weights to the second objective function, the weighted crowding distance Dist of each individual is calculated. p :

[0018] Among them, W safe The stability-weighted weight has a value greater than the time weight W. time f1 max f1 min These represent the maximum and minimum values ​​of all individuals within the current non-dominated hierarchy on the first objective function, respectively; f2 max f2 min These represent the maximum and minimum values ​​of all individuals within the current non-dominated hierarchy on the second objective function, respectively.

[0019] Furthermore, the feasibility-priority tournament selection specifically includes: Calculate the violation degree CV(p): Based on the established safety tilt angle hard constraint, calculate the individual's offset relative to the safety tilt angle limit.

[0020] Where, θ p , p These represent the lateral and longitudinal angles corresponding to individual p, respectively. Implement the feasibility comparison rule: Randomly select two individuals from the population for comparison. In the binary tournament selection, the comparison rules are as follows: If CV(p1) = 0 and CV(p2) > 0, then choose p1; If CV(p1) > 0 and CV(p2) > 0, then choose the one with the smaller CV value; If CV(p1)=CV(p2)=0, then a regular selection is performed based on the determined non-dominated sorting level and weighted crowding distance.

[0021] Furthermore, the adaptive genetic operator and physical boundary check specifically include: Adaptive mutation strategy: Based on the current generation t and the maximum number of iterations G of the population max The ratio of the mutation probability P is dynamically adjusted. m The system monitors the fitness distribution of individuals in the population in real time to determine population diversity. When the proportion of non-dominant individuals in the population is lower than a preset threshold or the average weighted crowding distance between adjacent individuals is less than a preset distance threshold, it is determined that population diversity has decreased, and the mutation probability P is increased. m To enhance the population's exploration capabilities and break out of local optima; Physical boundary correction: For the new individuals generated after crossover and mutation operations following the adaptive mutation strategy adjustment, the adjustment limits of each ballast tank are checked according to the set physical capacity constraints. Specifically, the physical capacity constraints are... 0≤v i,0 +Δv i ≤V max,i If the adjusted water volume exceeds the preset range, the truncation operator will force it back to the physically permissible range. If the adjusted water storage volume v of the new individual i,0 +Δv i If the value is less than 0, the adjusted water volume is corrected to 0 using a truncation operator; if the adjusted water volume is greater than V... max,i Then, the adjusted water volume is corrected to V by the truncation operator. max,i If the adjusted water volume is between 0 and V max,i Within the specified range, the new individual remains unchanged; The new individuals after the physical boundary correction are considered as valid individuals and included in the offspring population; The optimal solution decision-making and closed-loop control specifically include: The output Pareto optimal ballast allocation scheme solution set is used as a candidate scheme, and the optimal solution X is automatically selected from the candidate schemes according to a preset strategy. opt The optimal solution is represented as a solution vector X. opt =[Δv1, Δv2, …, Δv i ]; According to the solution vector X opt Generate commands for ballast pump start / stop and valve opening, specifically including: Pump control command: Adjust the water volume Δv of each ballast tank according to the solution vector. i With the corresponding pump set flow rate Q i The ratio is used to set the running timer for the corresponding pump group in each ballast tank, and the pump group running time. ; Valve control command: Adjust the water volume Δv according to the water volume of each ballast tank. i If Δv i If Δv > 0, then open the sea valve and the suction valve of the i-th ballast tank; if Δv i =0, then keep the sea valve and the suction valve of the i-th ballast tank closed; if Δv i If the value is less than 0, open the discharge valve and sea passage valve of the i-th ballast tank. During the execution of the pump control command and the valve control command by the ballast pump and valve, the real-time tilt angle θ is adjusted according to the preset period T. real and real-time pitch angle real Sampling was performed to obtain the real-time monitoring tilt angle; The real-time monitored tilt angle is compared with the predicted tilt angle θ from the total azimuth dynamic moment balance mathematical model. calc , calc Compare the values ​​and calculate the deviation. When the deviation value exceeds the preset threshold δ th When |θ is satisfied real -θ calc |>δ th or | real - calc |>δ th If the recalculation is triggered, the multi-objective optimization model will be recalculated immediately, and new ballast pump start / stop and valve opening commands will be generated based on the new Pareto optimal solution set obtained from the recalculation, so as to perform dynamic compensation.

[0022] A ballast water regulation system for a fully swivel crane piling vessel, used to implement any one of the ballast water regulation methods for a fully swivel crane piling vessel, the system comprising: The coordinate system module is used to establish the ship's appendage coordinate system O-XYZ, with the intersection of the ship's baseline and the mid-longitudinal section as the X-axis and positive in the direction of the bow, the mid-transverse section as the Y-axis and positive in the direction of the port side, and the vertical axis upward as the Z-axis. The data acquisition module is used to synchronously acquire real-time data at a preset period T. The real-time data includes hoisting load data, ship floating status data, and ballast tank status data. The central optimization decision-making module, based on the principles of ship statics, constructs a set of nonlinear equilibrium equations including the longitudinal and transverse heel coupling effect and free surface correction, generating a full-rotation dynamic moment balance mathematical model. The nonlinear equilibrium equations are used to calculate the ballast moment required to maintain ship balance based on the lifting load data, ship buoyancy data, and ballast tank status data. Based on the ballast moment, ballast water regulation is transformed into a dual-objective optimization problem with strong physical constraints, constructing a multi-objective optimization model. This model includes a first objective function of minimizing the allocation time and a second objective function of maximizing stability reserve, and sets hard constraints on safety heel angle, physical capacity constraints, and unidirectional exchange logic constraints. The calculation of the first and second objective functions depends on the parameters in the nonlinear equilibrium equations. An improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model, outputting a Pareto optimal solution set. The initialization of the improved NSGA-II algorithm depends on the established set of nonlinear equilibrium equations, and its individual evaluation depends on the constructed first objective function, second objective function, and various constraints. The ballast drive execution module is used to convert the output Pareto optimal solution set into ballast pump start / stop and valve opening commands, and trigger the improved NSGA-II algorithm for dynamic compensation based on the deviation between the real-time monitored tilt angle and the model predicted tilt angle in the collected real-time data.

[0023] Compared with the prior art, the embodiments of the present invention achieve the following beneficial effects: This invention establishes a ship appendage coordinate system; simultaneously collects data on lifting loads, ship buoyancy, and ballast tank status; constructs a full-rotation dynamic moment balance mathematical model incorporating longitudinal and transverse heel coupling effects and free surface correction; transforms ballast water regulation into a strongly constrained bi-objective optimization model aiming for the shortest regulation time and maximum stability reserve; it employs an improved NSGA-II algorithm—using heuristic moment compensation initialization, safety margin weighted congestion distance calculation, and feasibility-first tournament selection—for iterative solution; and converts the Pareto optimal solution set into ballast pump start / stop and valve opening commands, triggering dynamic compensation based on the deviation between real-time monitored heel angle and model-predicted heel angle. This invention achieves rapid, accurate, and safe generation of ballast water solutions, thereby enabling high-response, high-precision, and inherently safe stability regulation of a full-rotation crane piling vessel in complex dynamic construction environments. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating a ballast water adjustment method and system for a fully rotating hoisting piling vessel provided in an embodiment of the present invention. Figure 2 This is a system architecture diagram of a ballast water adjustment method and system for a full-rotation crane piling vessel provided in an embodiment of the present invention; Figure 3 This is an improved NSGA-II algorithm optimization execution flowchart of a ballast water adjustment method and system for a full-rotation crane piling vessel provided by an embodiment of the present invention; Figure 4 This is a schematic diagram of the ballast tank of a fully rotating crane piling vessel, which is a ballast water adjustment method and system for a fully rotating crane piling vessel provided in an embodiment of the present invention. Figure 5 This is a sorting table of f1 function values ​​for a ballast water adjustment method and system for a full-rotation crane piling vessel provided in an embodiment of the present invention; Figure 6 This is a table showing the comparison results of two randomly selected individuals in a tournament, provided by an embodiment of the present invention, regarding a method and system for adjusting ballast water on a fully rotating hoisting piling vessel. Among them, 1-ballast tank inlet, 2-electric remote control butterfly valve, 3-ballast tank assembly of the full-rotation crane piling vessel. Detailed Implementation

[0025] The present invention will now be described in detail with reference to the accompanying drawings.

[0026] Example 1

[0027] This invention provides a method for adjusting ballast water on a fully rotating hoisting piling vessel, referring to... Figures 1 to 4 ,include: S1. Establish the ship's appendage coordinate system O-XYZ, with the intersection of the ship's baseline and the mid-longitudinal section as the X-axis (positive towards the bow) and positive towards the bow, the mid-transverse section as the Y-axis (positive towards the port) and positive towards the port, and the vertical line upwards from the baseline as the Z-axis. S2. Real-time data is collected synchronously at a preset period T. The real-time data includes hoisting load data, ship floating status data, and ballast tank status data. S3. Based on the principle of ship statics, a set of nonlinear equilibrium equations including the longitudinal and transverse heel coupling effect and free surface correction is constructed to generate a full-rotation dynamic moment balance mathematical model. The set of nonlinear equilibrium equations is used to calculate the ballast moment required to maintain the balance of the ship based on the lifting load data, the ship's floating state data and the ballast tank state data, and in combination with the coordinate parameters in the ship's appendage coordinate system O-XYZ. S4. Based on the ballast moment, the ballast water regulation is transformed into a dual-objective optimization problem with strong physical constraints. A multi-objective optimization model is constructed. The multi-objective optimization model includes a first objective function of minimizing the allocation time and a second objective function of maximizing the stability reserve. Hard constraints on the safety tilt angle, physical capacity constraints, and unidirectional exchange logic constraints are set. The calculation of the first objective function and the second objective function depends on the parameters in the nonlinear equilibrium equation set. S5. The improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model and output the Pareto optimal solution set; wherein, the initialization of the improved NSGA-II algorithm depends on the established nonlinear equilibrium equation system, and its individual evaluation depends on the constructed first objective function, second objective function and various constraints. S6. Convert the output Pareto optimal solution set into ballast pump start / stop and valve opening commands, and trigger the improved NSGA-II algorithm for dynamic compensation based on the deviation between the real-time monitored tilt angle and the model predicted tilt angle in the collected real-time data.

[0028] Specifically, this embodiment deeply couples the physical characteristics of hoisting and piling operations with an improved multi-objective optimization algorithm, resulting in significant benefits: breaking through the real-time response bottleneck and achieving a synergistic improvement in response time and engineering execution efficiency. This embodiment uses a heuristic torque compensation initialization strategy, which replaces the blind search of traditional algorithms with prior physical knowledge, and directly anchors the initial population near the physical feasible region, significantly shortening the evolution process. At the same time, in conjunction with the unidirectional sea passage exchange logic constraint, it abandons the lagging and complex inter-cabin intermodulation mode in the traditional scheme. This combined optimization enables the algorithm to meet the requirements of rapid optimization while simplifying the control chain of the hardware execution end, ensuring that the ballast system can compensate for the instantaneous overturning moment generated by the dynamic full rotation of the piling frame in real time and efficiently. A red-line decision-making mechanism is constructed to ensure the inherent safety of piling operations under strong constraints. This embodiment introduces a feasibility-first tournament selection rule at the algorithm's underlying level, setting safety red lines such as 3.5° heel and 2° trim as rigid conditions for selecting individuals. This ensures that each generated instruction set inherently possesses safety. Combined with a safety margin-weighted congestion distance operator, the system can automatically identify and prioritize solutions with greater stability margins within the Pareto optimal solution set. This multi-criteria decision-making mechanism enables fully slewing crane piling vessels, such as... Figure 4 As shown, when faced with sudden sea conditions or extreme lifting conditions, it can not only output a solution, but also the safest solution, thus avoiding the risk of ship capsizing from the source of control logic. By coupling the physical model, precise attitude control of the entire rotational dynamic process is achieved. This embodiment establishes a six-degree-of-freedom moment balance model that includes the longitudinal and transverse tilt coupling effect, and introduces a key free liquid surface correction term. By quantitatively compensating for the stability reduction caused by the liquid sloshing in the ballast tank, this embodiment effectively corrects the prediction deviation caused by traditional decoupling calculation. This high-precision physical mapping enables the water volume adjustment scheme output by the system to accurately match the amplitude and rotation trajectory of the frame, solving the common problems of "insufficient adjustment" or "over-adjustment" in engineering practice, and improving the attitude stability during the pile driving alignment process.

[0029] In the above embodiments, specifically, the lifting load data includes the mass of the lifted object Pt, the current lifting radius R, and the real-time slewing angle α of the crane relative to the bow of the ship; these three factors together determine the magnitude and direction of the overturning moment generated by the lifting operation under the current working conditions. For example, when a piling vessel lifts a 50-ton pile with a boom radius of 30 meters and a slewing angle of 45°, the lateral overturning moment generated by the lifted object is Pt·g·R·sinα. The ship buoyancy data includes the ship's current average draft T. m (The average draft of the ship at the bow and stern in still water, used to calculate the ship's displacement and stability parameters), instantaneous heel angle θ real and pitch angle φ realThe ballast tank status data includes the current water volume set V0 = [v...] of all n ballast tanks on the ship. {1,0} , v {2,0} , ..., v {n,0} The current water volume set V0 is used to calculate the ship's current total displacement.

[0030] It should be noted that by synchronously collecting these three types of data at a preset period T, a real-time data foundation for ballast water regulation was established. Lifting load data reflects changes in dynamic construction conditions, ship buoyancy data reflects the current attitude state, and ballast tank status data reflects adjustable resources. The synchronous collection of these three types of data ensures the time consistency of parameters in subsequent mathematical model calculations, avoiding model errors caused by data asynchrony, and providing reliable data support for high-precision ballast water regulation.

[0031] In the above embodiments, specifically, the nonlinear equilibrium equation set includes: Lateral moment balance equation:

[0032] Longitudinal moment balance equation:

[0033] in, This refers to the ship's current total displacement. ;Δ light This indicates the empty ship's displacement, and the current water volume v i,0 This data originates from the ballast tank status data; , These are the horizontal and vertical centering heights, respectively. The free surface correction value is given, KG is the height of the ship's center of gravity, and θ is the free surface correction value. These are the rake angle and the pitch angle, respectively. , The moment of inertia is the surface area of ​​the liquid inside the chamber; This is the water volume adjustment amount for the i-th ballast tank (positive for water injection, negative for water drainage). , Let be the volume center coordinates of the i-th ballast tank in the ship's appendage coordinate system O-XYZ. For irregularly shaped ballast tanks, , These two values ​​are functions of the water depth inside the cabin, and can be looked up in real time through the cabin capacity table. ρ is the density of seawater, and g is the acceleration due to gravity.

[0034] It should be noted that the following are the current operating conditions of a piling vessel: load mass Pt = 80 tons, lifting radius R = 25 meters, slewing angle α = 60°, current displacement Δ = 5000 tons, lateral epicenter height KMT = 8.2 meters, center of gravity height KG = 5.5 meters, and free surface correction δGM. f =0.3 meters; The left side of the lateral moment balance equation (restoring moment): 5000×9.8×(8.2-5.5-0.3)×tanθ=5000×9.8×2.4×tanθ Right side (overturning moment + ballast adjustment moment): 80×9.8×25×sin60 ∘ +∑ρ·g·Δv i ·y i =16973.6×9.8+∑ρ·g·Δv i ·y i Using the above model, the solution can be obtained for a given Δv. i The equilibrium heel angle θ of the vessel, or conversely, the Δv required to achieve the target heel angle. i The model, for the first time, simultaneously considers the heel-roll coupling effect and free surface correction in ballast adjustment of piling vessels. The heel-roll coupling ensures the accuracy of moment calculation when both heel and roll are present. The free surface correction avoids overestimating the contribution of some loading compartments to stability. Compared with the traditional simplified model that ignores the effects of coupling and free surface, the calculation accuracy of the model is improved by more than 15%, providing an accurate physical basis for subsequent optimization.

[0035] In the above embodiments, specifically, the multi-objective optimization model includes: The first objective function is to minimize the allocation time (efficiency objective f1), which involves minimizing the maximum single pump operation time. The total time for ship-wide adjustment depends on the slowest pump. Assuming all participating ballast pumps operate in parallel, the objective function is designed as a "minimum-maximum" problem.

[0036] Where Ω represents the set of ballast tanks involved in the regulation, i.e., Δv i ≠0 cabins; Q i Let η be the actual flow rate of the pump group corresponding to the i-th ballast tank. Considering the pump's characteristic curve, the flow rate is not constant when the head changes. A piecewise linearized function is used to fit the pump's flow rate characteristics. sys The volumetric efficiency coefficient of the pump set is given, taking into account head losses along the flow path caused by elbows and valves; t value This refers to the valve opening and closing response delay time. The first objective function is used to drive the multi-objective optimization model to find a solution that can fully utilize the parallel capabilities of multiple pumps and avoid the inefficient situation of "one pump being overworked while the others are idle".

[0037] The second objective function is to maximize the stability reserve, minimize the residual tilt angle, and suppress the free surface: , Where, θ obj , obj All are ideal target tilt angles; ω θ , These are all weighting coefficients; in full-rotation operations, ω is typically... θ > S i Let be the free surface area in the i-th ballast tank; λ be the penalty factor; X represent the decision variable vector; X should be defined as: X=[Δv1,Δv2,…,Δv n ], which is the vector formed by the water volume regulation of the n ballast tanks of the whole ship; The second objective function aims to ensure the ship's ability to withstand wind and waves after adjustment, with the goal of minimizing the residual heel angle and lowering the center of gravity as much as possible through ballast distribution.

[0038] Constraint settings: Set hard constraints for safety tilt angle, physical capacity constraints, and unidirectional switching logic constraints; The safety tilt angle hard constraints include: tilt angle θ ≤ 3.5°, and pitch angle... If the value is ≤2°, the solution that violates this constraint is an infeasible solution; The physical capacity constraints include: 0≤v i,0 +Δv i ≤V max,i Wherein v i,0 V is derived from the ballast tank status data. max,i The maximum capacity of the i-th ballast tank; The unidirectional exchange logic constraints include: Δv i ·Δv j ≤0 is used to prioritize the direct exchange of ballast water with the outside, without inter-cabin water exchange. This refers to the water volume adjustment amount for the j-th ballast tank; that is, within the same adjustment cycle, complex "tank-to-tank" water transfer should be avoided, and a unified "sea passage" logic should be adopted to simplify the complexity of the valve control system.

[0039] It should be noted that the above-mentioned dual-objective optimization model is the first to consider both regulation efficiency (time) and operational safety (stability) in ballast regulation of piling vessels. Compared with the traditional model that only pursues the minimum regulation amount, this model can generate a variety of Pareto front solutions, allowing operators to select the optimal solution based on actual working conditions (such as weather conditions and operational urgency). The unidirectional exchange logic constraint ensures that ballast water is directly exchanged with the outside, avoiding the complexity and time loss of pipeline switching caused by inter-cabin swapping, thus shortening the regulation time by 20% to 30%.

[0040] In the above embodiments, specifically, an improved NSGA-II algorithm is used, such as... Figure 3 As shown, the constructed multi-objective optimization model is iteratively solved to output the Pareto optimal solution set, including: S51. Heuristic Torque Compensation Initialization: Set the evolutionary generation counter t=0, set the population size to N, and the maximum number of iterations to G. max Based on the nonlinear equilibrium equations, the mass P of the lifted object, the current lifting radius R, and the real-time slewing angle α, the theoretical regulating water volume benchmark value of each ballast tank is derived in reverse, and the initial parent population P0 is generated based on the theoretical regulating water volume benchmark value. S52. Non-dominated sorting and safety margin weighted crowding distance calculation: For each individual in the population, evaluate it according to the first objective function and the second objective function, calculate its allocation time objective function value and stability reserve objective function value, and then execute the stratification mechanism; S53. Feasibility-first tournament selection: Select outstanding individuals from the parent population to enter the pairing pool, define the constraint violation degree CV(p) of individual p according to the hard constraint of the safety tilt angle, and execute the feasibility comparison rule. S54. Adaptive Genetic Operators and Physical Boundary Checks: Perform crossover and mutation operations on individuals entering the pairing pool to produce the offspring population Q. t And perform boundary correction on the new individual based on the physical capacity constraints; S55. Population Merging and Elite Preservation Strategy: Merge the current parent population P... t With the generated offspring population Q t Merge them to form a combined population R t For all individuals in the combined population, the non-dominated ranking and safety margin weighted crowding distance calculation described in S52 are re-executed, and individuals are selected sequentially in ascending order of non-dominated level to fill the new population P. t+1 ; S56. Iteration Termination Judgment and Result Output: Check whether the current generation t has reached the maximum number of iterations G. maxIf the target is not reached, set t=t+1 and return to S53 to continue iterating. If the target is reached, output the individual with the first level of non-dominance as the solution set of the Pareto optimal ballast allocation scheme.

[0041] It should be noted that, compared with the standard NSGA-II, the improved algorithm in this embodiment has the following advantages: heuristic initialization increases the proportion of initial feasible solutions from less than 10% to more than 60%, significantly reducing invalid searches; combining safety margin weighted crowding distance and feasibility priority selection, the algorithm can converge within 50 generations, which is about 40% faster than the standard NSGA-II; the Pareto front distribution is more uniform, and both extremely safe and extremely efficient solutions can be effectively explored.

[0042] In the above embodiments, specifically, the heuristic torque compensation initialization includes: Calculate the theoretical compensation water volume: Using the lateral moment balance equation and the longitudinal moment balance equation, and based on the lifted object mass P, the current lifting radius R, and the real-time slewing angle α, deduce in reverse the overturning moment M generated by the current slewing. req The theoretical regulating water volume benchmark value required for each ballast tank (the overturning moment generated by the current lifting operation) :

[0043] Where β is the proportional gain coefficient, y i Let y be the Y-axis coordinate of the i-th ballast tank in the ship's appendage coordinate system O-XYZ; j Let be the Y-axis coordinate of the j-th ballast tank in the ship's appendage coordinate system O-XYZ; Generate seed individuals: Centered on the theoretical water regulation baseline value, a random perturbation term ε following a normal distribution N(0,σ²) is superimposed to generate heuristic seed individuals, which account for 30%~50% of the initial population; Population construction: The remaining individuals are randomly generated within the range that satisfies the set physical capacity constraints, and together with the heuristic seed individuals, they form the initial parent population P0.

[0044] It should be noted that, assuming the ship has four ballast tanks with Y-axis coordinates of: Tank 1: y1=+8m, Tank 2: y2=+4m, Tank 3: y3=-4m, Tank 4: y4=-8m, and the current overturning moment M... req =5000kN·m, β=1.0, ρ=1.025t / m 3 ; Theoretical adjustable water volume for each compartment: Cabin 1: Δv1 seed =(5000×8) / 164=243.9m3 Cabin 2: Δv2 seed =(5000×4) / 164=121.95m 3 Cabin 3: Δv3 seed =(5000×-4) / 164=-121.95m 3 Cabin 4: Δv4 seed =(5000×-8) / 164=-243.9m 3 It is evident that the water volume allocated to the remote compartments (compartments 1 and 4) is greater, reflecting the efficiency optimization of torque compensation.

[0045] In the above embodiments, specifically, the non-dominated sorting and safety margin weighted crowding distance calculation includes: Fast non-dominated sorting: Based on the dominance relationship between individuals, the population is divided into different non-dominated levels, where the individuals in the first non-dominated level are the current optimal solution set; Calculating the weighted crowding distance with a safety margin: Within the same non-dominated hierarchy, by introducing a safety weighting factor and assigning weights to the second objective function, the weighted crowding distance Dist of each individual is calculated. p :

[0046] Among them, W safe The stability-weighted weight has a value greater than the time weight W. time f1 and f2 are used to preferentially retain solutions with larger stability reserves when the settling times of the first objective function are similar, where f1 and f2 represent the first and second objective functions of the multi-objective optimization model, respectively; f1 max f1 min These represent the maximum and minimum values ​​of all individuals within the current non-dominated hierarchy on the first objective function, respectively; f2 max f2 min These represent the maximum and minimum values ​​of all individuals within the current non-dominated hierarchy on the second objective function, respectively.

[0047] It should be noted that, assuming there are 5 individuals in the same non-dominated hierarchy, sorted by f1 value as follows: Figure 5As shown, the crowding distances of boundary individuals A and E are set to infinity and are preferentially retained. Among the internal individuals, C has the largest crowding distance (0.285) and is most likely to be retained during population truncation. After introducing a safety weighting factor, the algorithm assigns a higher priority to stability objectives in terms of selection pressure. When the settling times of two schemes are similar, the algorithm will force the retention of the solution with a larger stability reserve, reflecting the engineering principle of "safety over efficiency". Compared with the equal-weighted crowding distance of the standard NSGA-II, this embodiment can generate a safer Pareto front, meeting the inherent safety requirements of piling vessel operations.

[0048] In the above embodiments, specifically, the feasibility-priority tournament selection includes: Calculate the violation degree CV(p): Based on the established safety tilt angle hard constraint, calculate the individual's offset relative to the safety tilt angle limit.

[0049] Where, θ p , p These represent the lateral and longitudinal angles corresponding to individual p, respectively. Implement the feasibility comparison rule: Randomly select two individuals from the population for comparison. In the binary tournament selection, the comparison rules are as follows: If CV(p1) = 0 and CV(p2) > 0, then choose p1; If CV(p1) > 0 and CV(p2) > 0, then choose the one with the smaller CV value; If CV(p1)=CV(p2)=0, then a regular selection is performed based on the determined non-dominated sorting level and weighted crowding distance.

[0050] It should be noted that, assuming the tournament randomly selects two individuals for comparison, such as... Figure 6 As shown, this selection mechanism prioritizes the satisfaction of safety constraints over the optimization objective, which aligns with the "safety first" principle in marine engineering. By explicitly handling infeasible solutions, the algorithm can effectively explore the feasible region boundary and avoid blindly searching for infeasible areas. Compared to the standard NSGA-II method of using penalty functions to handle constraints, this embodiment avoids the difficulty of adjusting penalty factors and ensures that the final solution set 100% satisfies the safety inclination angle hard constraint.

[0051] In the above embodiments, specifically, the adaptive genetic operator and physical boundary check include: Adaptive mutation strategy: Based on the current generation t and the maximum number of iterations G of the population max The ratio of the mutation probability P is dynamically adjusted. mThe system monitors the fitness distribution of individuals in the population in real time to determine population diversity. When the proportion of non-dominant individuals in the population is lower than a preset threshold or the average weighted crowding distance between adjacent individuals is less than a preset distance threshold, it is determined that population diversity has decreased, and the mutation probability P is increased. m To enhance the population's exploration capabilities and break out of local optima; Physical boundary correction: For the new individuals generated after crossover and mutation operations following the adaptive mutation strategy adjustment, the adjustment limits of each ballast tank are checked according to the set physical capacity constraints. Specifically, the physical capacity constraints are... 0≤v i,0 +Δv i ≤V max,i If the adjusted water volume exceeds the preset range, the truncation operator will force it back to the physically permissible range. If the adjusted water storage volume v of the new individual i,0 +Δv i If the value is less than 0, the adjusted water volume is corrected to 0 using a truncation operator; if the adjusted water volume is greater than V... max,i Then, the adjusted water volume is corrected to V by the truncation operator. max,i If the adjusted water volume is between 0 and V max,i Within the specified range, the new individual remains unchanged; The new individuals after the physical boundary correction are considered as valid individuals and included in the offspring population; The optimal solution decision-making and closed-loop control specifically include: The output Pareto optimal ballast allocation scheme solution set is used as a candidate scheme, and the optimal solution X is automatically selected from the candidate schemes according to a preset strategy. opt The optimal solution is represented as a solution vector X. opt =[Δv1, Δv2, …, Δv i ]; According to the solution vector X opt Generate commands for ballast pump start / stop and valve opening, specifically including: Pump control command: Adjust the water volume Δv of each ballast tank according to the solution vector. i With the corresponding pump set flow rate Q i The ratio is used to set the running timer for the corresponding pump group in each ballast tank, and the pump group running time. ; Valve control command: Adjust the water volume Δv according to the water volume of each ballast tank. i If Δv i If Δv > 0, then open the sea valve and the suction valve of the i-th ballast tank; if Δv i=0, then keep the sea valve and the suction valve of the i-th ballast tank closed; if Δv i If the value is less than 0, open the discharge valve and sea passage valve of the i-th ballast tank. During the execution of the pump control command and the valve control command by the ballast pump and valve, the real-time tilt angle θ is adjusted according to the preset period T. real and real-time pitch angle real Sampling was performed to obtain the real-time monitoring tilt angle; The real-time monitored tilt angle is compared with the predicted tilt angle θ from the total azimuth dynamic moment balance mathematical model. calc , calc Compare the values ​​and calculate the deviation. When the deviation value exceeds the preset threshold δ th When |θ is satisfied real -θ calc |>δ th or | real - calc |>δ th If the recalculation is triggered, the multi-objective optimization model will be recalculated immediately, and new ballast pump start / stop and valve opening commands will be generated based on the new Pareto optimal solution set obtained from the recalculation, so as to perform dynamic compensation.

[0052] Example 2

[0053] A ballast water regulation system for a fully rotating hoisting piling vessel, such as Figure 2 As shown, the system includes: The coordinate system module is used to establish the ship's appendage coordinate system O-XYZ, with the intersection of the ship's baseline and the mid-longitudinal section as the X-axis and positive in the direction of the bow, the mid-transverse section as the Y-axis and positive in the direction of the port side, and the vertical axis upward as the Z-axis. The data acquisition module is used to synchronously acquire real-time data at a preset period T. The real-time data includes hoisting load data, ship floating status data, and ballast tank status data. The central optimization decision-making module, based on the principles of ship statics, constructs a set of nonlinear equilibrium equations including the longitudinal and transverse heel coupling effect and free surface correction, generating a full-rotation dynamic moment balance mathematical model. The nonlinear equilibrium equations are used to calculate the ballast moment required to maintain ship balance based on the lifting load data, ship buoyancy data, and ballast tank status data. Based on the ballast moment, ballast water regulation is transformed into a dual-objective optimization problem with strong physical constraints, constructing a multi-objective optimization model. This model includes a first objective function of minimizing the allocation time and a second objective function of maximizing stability reserve, and sets hard constraints on safety heel angle, physical capacity constraints, and unidirectional exchange logic constraints. The calculation of the first and second objective functions depends on the parameters in the nonlinear equilibrium equations. An improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model, outputting a Pareto optimal solution set. The initialization of the improved NSGA-II algorithm depends on the established set of nonlinear equilibrium equations, and its individual evaluation depends on the constructed first objective function, second objective function, and various constraints. The ballast drive execution module is used to convert the output Pareto optimal solution set into ballast pump start / stop and valve opening commands, and trigger the improved NSGA-II algorithm for dynamic compensation based on the deviation between the real-time monitored tilt angle and the model predicted tilt angle in the collected real-time data.

[0054] It should be understood that the above embodiments are one or more embodiments of the present invention, and there are many other embodiments and variations based on the present invention; any variations and modifications made by those skilled in the art through the present invention without making pioneering innovations are all within the protection scope of the present invention.

Claims

1. A method for adjusting ballast water on a fully rotating hoisting piling vessel, characterized in that, The method includes: Establish the ship's appendage coordinate system O-XYZ, with the intersection of the ship's baseline and the mid-longitudinal section as the X-axis and positive in the direction of the bow, the mid-transverse section as the Y-axis and positive in the direction of the port side, and the vertical axis upward as the Z-axis. Real-time data is collected synchronously at a preset period T, including hoisting load data, ship floating status data, and ballast tank status data. Based on the principles of ship statics, a set of nonlinear equilibrium equations, including the longitudinal and transverse heel coupling effect and free surface correction, is constructed to generate a full-rotation dynamic moment balance mathematical model. The set of nonlinear equilibrium equations is used to calculate the ballast moment required to maintain the ship's balance based on the lifting load data, the ship's floating state data, and the ballast tank state data, combined with the coordinate parameters in the ship's appendage coordinate system O-XYZ. Based on the ballast moment, the ballast water regulation is transformed into a bi-objective optimization problem with strong physical constraints. A multi-objective optimization model is constructed, which includes a first objective function of minimizing the allocation time and a second objective function of maximizing the stability reserve. Hard constraints on the safety tilt angle, physical capacity constraints, and unidirectional exchange logic constraints are set. The calculation of the first objective function and the second objective function depends on the parameters in the nonlinear equilibrium equation system. An improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model and output the Pareto optimal solution set. The initialization of the improved NSGA-II algorithm depends on the established nonlinear equilibrium equation system, and its individual evaluation depends on the constructed first objective function, second objective function, and various constraints. The output Pareto optimal solution set is converted into ballast pump start / stop and valve opening commands, and the improved NSGA-II algorithm is triggered to perform dynamic compensation based on the deviation between the real-time monitored tilt angle and the model predicted tilt angle in the collected real-time data.

2. The ballast water adjustment method for a full-rotation crane piling vessel according to claim 1, characterized in that, The lifting load data includes the mass of the lifted object Pt, the current lifting radius R, and the real-time slewing angle α of the crane relative to the bow of the ship; the ship's buoyancy data includes the ship's current average draft T. m Instantaneous tilt angle θ real and pitch angle φ real The ballast tank status data includes the current water volume set V0 = [v...] of all n ballast tanks on the ship. {1,0} , v {2,0} , ..., v {n,0} The current water volume set V0 is used to calculate the ship's current total displacement.

3. The ballast water adjustment method for a full-rotation crane piling vessel according to claim 2, characterized in that, The nonlinear equilibrium equation set specifically includes: Lateral moment balance equation: Longitudinal moment balance equation: in, This refers to the ship's current total displacement. ;Δ light This indicates the empty ship's displacement, and the current water volume v i,0 This data originates from the ballast tank status data; , These are the horizontal and vertical centering heights, respectively. The free surface correction value is given, KG is the height of the ship's center of gravity, and θ is the free surface correction value. These are the rake angle and the pitch angle, respectively. , The moment of inertia is the surface area of ​​the liquid inside the chamber; This refers to the water volume adjustment amount for the i-th ballast tank; , Let ρ be the volume center coordinate of the i-th ballast tank in the ship's appendage coordinate system O-XYZ, where ρ is the density of seawater and g is the acceleration due to gravity.

4. The ballast water adjustment method for a full-rotation crane piling vessel according to claim 3, characterized in that, The multi-objective optimization model includes: The primary objective function is to minimize the deployment time and the maximum single-pump operating time. Where Ω represents the set of ballast tanks involved in the regulation, i.e., Δv i ≠0 cabins; Q i η represents the actual flow rate of the pump unit corresponding to the i-th ballast tank. sys t is the volumetric efficiency coefficient of the pump set. value f1(·) represents the valve opening and closing response delay time; f1(·) represents the first objective function; The second objective function is to maximize the stability reserve, minimize the residual tilt angle, and suppress the free surface: ， Where, θ obj , obj All are ideal target tilt angles; ω θ , All are weighting coefficients; S i Let be the free surface area in the i-th ballast tank; λ be the penalty factor; X be the decision variable vector; f2(·) be the second objective function; Constraint settings: Set hard constraints for safety tilt angle, physical capacity constraints, and unidirectional switching logic constraints; The safety tilt angle hard constraints include: tilt angle θ ≤ 3.5°, and pitch angle... If the value is ≤2°, the solution that violates this constraint is an infeasible solution; The physical capacity constraints include: 0≤v i,0 +Δv i ≤V max,i Wherein v i,0 V is derived from the ballast tank status data. max,i The maximum capacity of the i-th ballast tank; The unidirectional exchange logic constraints include: Δv i ·Δv j ≤0 is used to prioritize the direct exchange of ballast water with the outside, without inter-cabin water exchange. This represents the water volume adjustment amount for the j-th ballast tank.

5. A method for adjusting ballast water on a fully rotating hoisting piling vessel according to claim 4, characterized in that, The improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model, outputting a Pareto optimal solution set, including: S51. Heuristic Torque Compensation Initialization: Set the evolutionary generation counter t=0, set the population size to N, and the maximum number of iterations to G. max Based on the nonlinear equilibrium equations, the mass P of the lifted object, the current lifting radius R, and the real-time slewing angle α, the theoretical regulating water volume benchmark value of each ballast tank is derived in reverse, and the initial parent population P0 is generated based on the theoretical regulating water volume benchmark value. S52. Non-dominated sorting and safety margin weighted crowding distance calculation: For each individual in the population, evaluate it according to the first objective function and the second objective function, calculate its allocation time objective function value and stability reserve objective function value, and then execute the stratification mechanism; S53. Feasibility-first tournament selection: Select outstanding individuals from the parent population to enter the pairing pool, define the constraint violation degree CV(p) of individual p according to the hard constraint of the safety tilt angle, and execute the feasibility comparison rule. S54. Adaptive Genetic Operators and Physical Boundary Checks: Perform crossover and mutation operations on individuals entering the pairing pool to produce the offspring population Q. t And perform boundary correction on the new individual based on the physical capacity constraints; S55. Population Merging and Elite Preservation Strategy: Merge the current parent population P... t With the generated offspring population Q t Merge them to form a combined population R t For all individuals in the combined population, the non-dominated ranking and safety margin weighted crowding distance calculation described in S52 are re-executed, and individuals are selected sequentially in ascending order of non-dominated level to fill the new population P. t+1 ; S56. Iteration Termination Judgment and Result Output: Check whether the current generation t has reached the maximum number of iterations G. max If the target is not reached, set t=t+1 and return to S53 to continue iterating. If the target is reached, output the individual with the first level of non-dominance as the solution set of the Pareto optimal ballast allocation scheme.

6. The ballast water adjustment method for a full-rotation crane piling vessel according to claim 5, characterized in that, The heuristic torque compensation initialization includes: Calculate the theoretical compensation water volume: Using the lateral moment balance equation and the longitudinal moment balance equation, and based on the lifted object mass P, the current lifting radius R, and the real-time slewing angle α, deduce in reverse the overturning moment M generated by the current slewing. req Required theoretical regulating water volume benchmark values ​​for each ballast tank : Where β is the proportional gain coefficient, y i Let y be the Y-axis coordinate of the i-th ballast tank in the ship's appendage coordinate system O-XYZ; j Let be the Y-axis coordinate of the j-th ballast tank in the ship's appendage coordinate system O-XYZ; Generate seed individuals: Centered on the theoretical water regulation baseline value, a random perturbation term ε following a normal distribution N(0,σ²) is superimposed to generate heuristic seed individuals, which account for 30%~50% of the initial population; Population construction: The remaining individuals are randomly generated within the range that satisfies the set physical capacity constraints, and together with the heuristic seed individuals, they form the initial parent population P0.

7. The ballast water adjustment method for a full-rotation crane piling vessel according to claim 5, characterized in that, The non-dominated sorting and safety margin weighted crowd distance calculation specifically includes: Fast non-dominated sorting: Based on the dominance relationship between individuals, the population is divided into different non-dominated levels, where the individuals in the first non-dominated level are the current optimal solution set; Calculating the weighted crowding distance with a safety margin: Within the same non-dominated hierarchy, by introducing a safety weighting factor and assigning weights to the second objective function, the weighted crowding distance Dist of each individual is calculated. p : Among them, W safe The stability-weighted weight has a value greater than the time weight W. time f1 max f1 min These represent the maximum and minimum values ​​of all individuals within the current non-dominated hierarchy on the first objective function, respectively; f2 max f2 min These represent the maximum and minimum values ​​of all individuals within the current non-dominated hierarchy on the second objective function, respectively.

8. A method for adjusting ballast water on a fully rotating hoisting piling vessel according to claim 5, characterized in that, The feasibility-priority tournament selection specifically includes: Calculate the violation degree CV(p): Based on the established safety tilt angle hard constraint, calculate the individual's offset relative to the safety tilt angle limit. Where, θ p , p These represent the lateral and longitudinal angles corresponding to individual p, respectively. Implement the feasibility comparison rule: Randomly select two individuals from the population for comparison. In the binary tournament selection, the comparison rules are as follows: If CV(p1) = 0 and CV(p2) > 0, then choose p1; If CV(p1) > 0 and CV(p2) > 0, then choose the one with the smaller CV value; If CV(p1)=CV(p2)=0, then a regular selection is performed based on the determined non-dominated sorting level and weighted crowding distance.

9. A method for adjusting ballast water on a fully rotating hoisting piling vessel according to claim 5, characterized in that, The adaptive genetic operator and physical boundary check specifically include: Adaptive mutation strategy: Based on the current generation t and the maximum number of iterations G of the population max The ratio of the mutation probability P is dynamically adjusted. m The system monitors the fitness distribution of individuals in the population in real time to determine population diversity. When the proportion of non-dominant individuals in the population is lower than a preset threshold or the average weighted crowding distance between adjacent individuals is less than a preset distance threshold, it is determined that population diversity has decreased, and the mutation probability P is increased. m To enhance the population's exploration capabilities and break out of local optima; Physical boundary correction: For the new individuals generated by crossover and mutation operations after adjustment using the adaptive mutation strategy, the adjustment limits of each ballast tank are checked according to the set physical capacity constraints, specifically 0 ≤ v. i,0 +Δv i ≤V max,i If the adjusted water volume exceeds the preset range, the truncation operator will force it back to the physically permissible range. If the adjusted water storage volume v of the new individual i,0 +Δv i If the value is less than 0, the adjusted water volume is corrected to 0 using a truncation operator; if the adjusted water volume is greater than V... max,i Then, the adjusted water volume is corrected to V by the truncation operator. max,i If the adjusted water volume is between 0 and V max,i Within the specified range, the new individual remains unchanged; The new individuals after the physical boundary correction are considered as valid individuals and included in the offspring population; The optimal solution decision-making and closed-loop control specifically include: The output Pareto optimal ballast allocation scheme solution set is used as a candidate scheme, and the optimal solution X is automatically selected from the candidate schemes according to a preset strategy. opt The optimal solution is represented as a solution vector X. opt =[Δv1, Δv2, …, Δv i ]; According to the solution vector X opt Generate commands for ballast pump start / stop and valve opening, specifically including: Pump control command: Adjust the water volume Δv of each ballast tank according to the solution vector. i With the corresponding pump set flow rate Q i The ratio is used to set the running timer for the corresponding pump group in each ballast tank, and the pump group running time. ； Valve control command: Adjust the water volume Δv according to the water volume of each ballast tank. i If Δv i If Δv > 0, then open the sea valve and the suction valve of the i-th ballast tank; if Δv i =0, then keep the sea valve and the suction valve of the i-th ballast tank closed; if Δv i If the value is less than 0, open the discharge valve and sea passage valve of the i-th ballast tank. During the execution of the pump control command and the valve control command by the ballast pump and valve, the real-time tilt angle θ is adjusted according to the preset period T. real and real-time pitch angle real Sampling was performed to obtain the real-time monitoring tilt angle; The real-time monitored tilt angle is compared with the predicted tilt angle θ from the total azimuth dynamic moment balance mathematical model. calc , calc Compare the values ​​and calculate the deviation. When the deviation value exceeds the preset threshold δ th When |θ is satisfied real -θ calc |>δ th or | real - calc |>δ th If the recalculation is triggered, the multi-objective optimization model will be recalculated immediately, and new ballast pump start / stop and valve opening commands will be generated based on the new Pareto optimal solution set obtained from the recalculation, so as to perform dynamic compensation.

10. A ballast water regulation system for a fully rotating crane-driven piling vessel, used to implement the ballast water regulation method for a fully rotating crane-driven piling vessel as described in any one of claims 1-9, characterized in that, The system includes: The coordinate system module is used to establish the ship's appendage coordinate system O-XYZ, with the intersection of the ship's baseline and the mid-longitudinal section as the X-axis and positive in the direction of the bow, the mid-transverse section as the Y-axis and positive in the direction of the port side, and the vertical axis upward as the Z-axis. The data acquisition module is used to synchronously acquire real-time data at a preset period T. The real-time data includes hoisting load data, ship floating status data, and ballast tank status data. The central optimization decision-making module, based on the principles of ship statics, constructs a set of nonlinear equilibrium equations including the longitudinal and transverse heel coupling effect and free surface correction, generating a full-rotation dynamic moment balance mathematical model. The nonlinear equilibrium equations are used to calculate the ballast moment required to maintain ship balance based on the lifting load data, ship buoyancy data, and ballast tank status data. Based on the ballast moment, ballast water regulation is transformed into a dual-objective optimization problem with strong physical constraints, constructing a multi-objective optimization model. This model includes a first objective function of minimizing the allocation time and a second objective function of maximizing stability reserve, and sets hard constraints on safety heel angle, physical capacity constraints, and unidirectional exchange logic constraints. The calculation of the first and second objective functions depends on the parameters in the nonlinear equilibrium equations. An improved NSGA-II algorithm is used to iteratively solve the constructed multi-objective optimization model, outputting a Pareto optimal solution set. The initialization of the improved NSGA-II algorithm depends on the established set of nonlinear equilibrium equations, and its individual evaluation depends on the constructed first objective function, second objective function, and various constraints. The ballast drive execution module is used to convert the output Pareto optimal solution set into ballast pump start / stop and valve opening commands, and trigger the improved NSGA-II algorithm for dynamic compensation based on the deviation between the real-time monitored tilt angle and the model predicted tilt angle in the collected real-time data.

Images