A method of ship power optimization

Abstract

The present application relates to the technical field of ship power control, in particular to a ship power optimization method. The present application comprises the steps of data acquisition and parameter initialization, setting the optimization variable constraint range, fitness calculation, speed and position updating, pulse parameter updating, convergence judgment and output of the optimal solution. The method of the present application solves the problems of weak global optimization ability, low local precision and poor dynamic adaptability of traditional optimization algorithms in ship power systems by fusing advantages and adapting to ship power characteristics, and provides an efficient and reliable intelligent decision-making tool for multi-objective optimization of hybrid power ships.

Description

Technical Field

[0001] This invention relates to the field of ship power control technology, and in particular to a method for ship power optimization. Background Technology

[0002] As the global shipping industry transforms towards green and intelligent technologies, ship propulsion systems are gradually evolving from traditional single-engine diesel drives to a multi-source hybrid power model combining diesel generators, new energy sources (batteries, hydrogen fuel, etc.), and auxiliary energy (wind power). This model, through the synergistic complementarity of multiple energy sources, demonstrates significant advantages in reducing fuel consumption and emissions, and improving range, and has become the mainstream development direction for inland waterway and coastal vessels. However, existing power control and optimization technologies for multi-source hybrid power vessels still have many shortcomings, making it difficult to meet the demands for efficient and safe operation in complex navigation scenarios. These shortcomings are specifically reflected in the following aspects:

[0003] Insufficient precision in power coordination control: Existing ships mostly rely on fixed control strategies (such as activating batteries under low load and switching to diesel generators under high load) for power distribution, without fully considering the dynamic changes in navigation status and environmental conditions. For example, the angle of attack adjustment during wind-assisted navigation lacks real-time linkage with the power output of the diesel generator set, and the wind energy utilization rate is only maintained at a basic level; the battery charging and discharging control is not combined with the operating efficiency curve of the diesel generator set, resulting in large power fluctuations during multi-energy coupled operation, frequent equipment start-ups and shutdowns, and shortened service life of the diesel generator set and batteries.

[0004] Poor adaptability of optimization algorithms: Traditional optimization algorithms (such as single particle swarm optimization and genetic algorithms) have obvious limitations in ship power optimization: either they have weak global optimization capabilities and are prone to getting stuck in local optima (such as pursuing the lowest fuel consumption while ignoring emissions and equipment stability); or they have insufficient local search accuracy and are difficult to adapt to complex scenarios with multiple coupled variables (such as the coordinated optimization of diesel generator power, battery SOC, wing angle of attack and propeller speed).

[0005] Incomplete consideration of constraints: Existing power optimization schemes do not fully take into account the physical limits and safety thresholds of ship equipment. The absence or careless handling of these constraints may lead to optimization results that exceed the safe operating range of the equipment, causing the risk of failure. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method for optimizing ship propulsion. This invention achieves efficient synergy among multiple energy sources through systematic parameter acquisition, intelligent algorithm optimization, and constraint control, thereby improving the economy, environmental friendliness, and safety of ship navigation.

[0007] The technical solution of this invention is: a method for optimizing ship dynamics, characterized by comprising the following steps:

[0008] Step 1: Data Acquisition and Parameter Initialization Steps

[0009] The data for the variables requiring optimization are standardized and parameters are set using the following formula:

[0010] xinorm represents normalized data. i represents the original data. The maximum value in the data. The minimum value in the data;

[0011] Step 2: Set the constraint range for the optimization variables:

[0012] Based on the different ship performance characteristics, optimize the value range of different parameters.

[0013] Step 3: Fitness Calculation

[0014] For each particle, the J value is calculated using the multi-objective function, and the formula is as follows:

[0015] minJ = ωy × A + ωy × B + ... + ωy × C, where y takes values ​​from 1 and 2 to the particle dimension; A, B, and C are the factor values, respectively.

[0016] Step 4: Speed ​​and Position Update

[0017] The new velocity vi(t+1) of each particle is calculated according to the velocity update formula; the formula is as follows: vi(t+1) = Ai(t)vi(t) + c1fi(xi(t)−gbest) + c2fi(xi(t)−pibest)

[0018] vi(t) is the velocity vector of the i-th particle at time t; c1 and c2 are acceleration factors, fi is the pulse frequency of the i-th particle, fi = fmin + (fmax - fmin) × rand(), fmin and fmax are the upper and lower limits of the frequency, rand() is a random number in (0, 1); xi(t) is the position vector of the i-th particle at time t, gbes is the global optimal position vector of all particles, and pibest is the individual optimal position vector of the i-th particle.

[0019] Calculate the new position: xi(t+1) = xi(t) + vi(t+1). If the new position exceeds the constraint range, then trim to the boundary.

[0020] Step 5: Pulse parameter update:

[0021] The updated formula calculates Ai(t+1) = α × Ai(t), where α is the attenuation coefficient.

[0022] Step 6: Convergence check

[0023] If the convergence condition is not met, return to step 3 and continue iterating; if the convergence condition is met, proceed to step 7.

[0024] Step 7: Output the optimal solution.

[0025] Output the combination of dynamic parameters corresponding to the globally optimal position gbest.

[0026] According to the above-described method for optimizing ship power, the optimization variables are: diesel engine power, battery, wing angle of attack, and propeller speed.

[0027] According to the above-mentioned method for optimizing ship power, the fitness calculation process is as follows: minJ=ω1×fuel+ω2×emission+ω3×speed+ω4×wind_efficiency, where ω1+ω2+ω3+ω4=1; Fuel is the diesel generator power factor value; Emission is the battery life factor value; speed is the propeller speed factor value; and wind_efficiency is the wind energy utilization factor value.

[0028] According to the above-mentioned method for ship dynamics optimization, the characteristics are: the parameters are set as follows: population size = 700, maximum number of iterations T = 300, fmin = 1, fmax = 5, and initial loudness Ai (0) = 1.0.

[0029] According to the above-described method for optimizing ship power, the characteristic feature is that the power of the diesel engine unit ranges from 80kW to 2000kW.

[0030] According to the above-described method for optimizing ship power, the characteristic feature is that the battery SOC ranges from 20% to 90%.

[0031] According to the above-described method for optimizing ship power, the characteristic feature is that the thruster speed ranges from 200 rpm to 1800 rpm.

[0032] According to the above-described method for optimizing ship dynamics, the characteristic feature is that the angle of attack of the winglet ranges from -15° to 30°.

[0033] According to the above-described method for ship dynamics optimization, the characteristic is that: c1=2−2t / T, c2=2t / T, t is the current iteration number, and T is the maximum iteration number.

[0034] According to the method for ship dynamics optimization described above, the characteristic is that: α=0.95, Ai(t) is greater than or equal to 0.01.

[0035] According to the above-described method for ship dynamics optimization, the convergence condition is that the global optimum value J fluctuates by less than 1% over 20 consecutive iterations, or the number of iterations reaches T=300.

[0036] The beneficial effects of this invention are: by integrating advantages and adapting to the characteristics of ship propulsion, the method of this invention solves the problems of "weak global optimization ability, low local accuracy and poor dynamic adaptability" of traditional optimization algorithms in ship propulsion systems, and provides an efficient and reliable intelligent decision-making tool for multi-objective optimization of hybrid ships. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0038] The present invention will now be further described with reference to the accompanying drawings:

[0039] Explanation of the name:

[0040] Diesel generator sets are the core power units of a ship's propulsion system. They consist of the diesel engine itself, auxiliary equipment, and control modules. By burning diesel fuel, they convert chemical energy into mechanical energy, providing stable power for ship navigation, power generation, and the operation of auxiliary equipment.

[0041] Electric power: Hydrogen fuel cells, lithium iron phosphate batteries, etc., are connected through a multi-source flexible DC ring network. The ring network uses solid-state switches to support bidirectional power flow. Battery SOC is the battery's "remaining capacity percentage," with a value ranging from 0% to 100%.

[0042] Wind power assistance: The angle of attack of the wind turbine blades is adjustable. By using the optimal angle of attack, the wind energy utilization rate can be improved, and energy recovery can be achieved in conjunction with the shaft-driven generator.

[0043] like Figure 1 As shown, the present invention provides a method for optimizing ship propulsion, applicable to the control and switching of multi-powered ship propulsion systems, comprising the following steps:

[0044] Step 1: Data Acquisition and Parameter Initialization Steps

[0045] The optimization variables (particle dimension) are determined based on various hull types: for example, in this invention, the optimization variables are selected as diesel engine power, battery, wing angle of attack, and propeller speed, forming a 4-dimensional variable. Different optimization variables can be selected depending on the hull's operating route; this invention takes hulls operating in rivers and coastal areas as examples.

[0046] Various parameters related to ship operation are collected, including data acquisition and preprocessing: The multi-dimensional sensing layer collects environmental parameters in real time, such as diesel engine speed (or propeller speed), battery SOC, and wing angle of attack. Ship navigation status data includes real-time ship speed (accuracy ±0.1 knots), heading (±1°), hull load (e.g., cargo load, ballast water status), propeller speed (±1 rpm), and propulsion power (±2% of rated power), collected in real-time via ship GPS, speed sensors, and torque sensors, with a sampling frequency of 1 Hz. Meteorological and sea state data includes wind speed (±0.5 m / s), wave height (±0.1 m), current speed (±0.2 m / s), and wind direction (±5°), collected via shipborne weather radar, wave sensors, and hydrological monitoring equipment, with a sampling frequency of 10 s / time. Powertrain performance data: fuel consumption rate (±0.5g / kWh), exhaust temperature (±2℃), lubricating oil temperature (±1℃), and battery pack SOC (±1%) and charge / discharge current (±5A), with a sampling frequency of 500ms / time.

[0047] The data for the variables requiring optimization are standardized using the following formula:

[0048] xinorm represents normalized data. i represents the original data. The maximum value in the data. It is the minimum value in the data.

[0049] The normalized variables are used as input for training. For example, (300kW, 50%, 1200rpm, 5°) is used as a set of data. The algorithm parameters are set as follows: population size = 700, maximum number of iterations T = 300, pulse frequency range (fmin = 1, fmax = 5), initial loudness Ai (0) = 1.0, and initial loudness vi (0) is set to the maximum response value of each parameter.

[0050] Step 2: Set the constraint range for the optimization variables:

[0051] Based on the different performance characteristics of various ships, the value ranges of different parameters are optimized. For example, the PDE value range for diesel engine power is 80kW to 2000kW; that is, below 80kW, the diesel engine is prone to "carbon buildup and stalling", and above 2000kW, it exceeds the design load and causes damage. The SOC value range for the battery is (20% to 90%); that is, below 20%, deep discharge of the battery shortens its lifespan, and above 90%, overcharging causes safety risks.

[0052] The propeller speed n ranges from 200 rpm to 1800 rpm. Below 200 rpm, the propeller is prone to stalling, and above 1800 rpm, it exceeds the shaft system's tolerance limit. The wing angle of attack ranges from -15° to 30°.

[0053] During the power switching process, the optimization range can be changed. For example, if the maximum power of the diesel unit drops due to a fault, its range can be adjusted to 80kW or 1000kW.

[0054] Step 3: Fitness Calculation

[0055] For each particle (combination of dynamic parameters), the J value is calculated by substituting it into a multi-objective function. The smaller the J value, the higher the fitness. In this embodiment, the multi-objective function is to balance fuel consumption, battery life optimization, and operating speed optimization.

[0056] minJ = ωy × A + ωy × B + ... + ωy × C, where y takes values ​​from 1 and 2 to the particle dimension; A, B, and C are the factor values.

[0057] In this embodiment, the particle dimension is 4, and the target factors are diesel generator power, battery life, propeller speed, wind energy utilization, etc. The method of the present invention can add or change factors according to actual needs, such as ship weight, wave height, etc. The present invention can meet the power optimization schemes of different scenarios or different needs through fitness calculation.

[0058] minJ=ω1×fuel+ω2×emission+ω3×speed+ω4×wind_efficiency, where ω1+ω2+ω3+ω4=1

[0059] Logic for setting target weights: Based on the core needs of ship operation (prioritizing economy, while taking into account environmental protection and safety), weights ω1=0.4, ω2=0.2, ω3=0.1, and ω4=0.3 are set to transform the multi-objective problem into a single-objective optimization.

[0060] Fuel is the power factor value of the diesel generator set; Emission is the battery life factor value; speed is the propeller speed factor value; wind_efficiency is the wind energy utilization factor value.

[0061] Record the individual optimal position pibest (the parameter corresponding to the minimum J in its own history) and the global optimal position gbest (the parameter corresponding to the minimum J in the entire population) for each particle.

[0062] Step 4: Speed ​​and Position Update

[0063] Calculate the new velocity vi(t+1) of each particle according to the velocity update formula;

[0064] Speed ​​update is the "core driving force" of algorithm optimization, which determines the movement direction and step size of the particle (optimization variable). The formula is as follows: vi(t+1)=Ai(t)vi(t)+c1fi(xi(t)−gbest)+c2fi(xi(t)−pibest)

[0065] vi(t) is the velocity vector of the i-th particle at time t; c1 and c2 are acceleration factors, dynamically adjusted: c1=2−2t / T, c2=2t / T (t is the current iteration number, T is the maximum iteration number); fi is the pulse frequency of the i-th particle, fi=fmin+(fmax−fmin)×rand() (fmin, fmax are the upper and lower limits of the frequency, rand() is a (0,1) random number); xi(t) is the position vector of the i-th particle at time t, gbes is the global optimal position vector of all particles, and pibest is the individual optimal position vector of the i-th particle.

[0066] Calculate the new position: xi(t+1) = xi(t) + vi(t+1). If the new position exceeds the constraint range, it is clipped to the boundary. For example, if SOC < 20%, it is clipped to the boundary (SOC = 20%). The method of this invention integrates boundary and update, reducing the number of optimizations and ensuring that the final predicted data matches the actual operation of the equipment.

[0067] Step 5: Pulse parameter update:

[0068] The formula for calculating Ai(t+1) is updated; the particle search step size is adjusted based on the new Ai; loudness is updated: Ai(t+1) = α × Ai(t) (α = 0.95, attenuation coefficient), and the range of Ai(t) can be set, such as Ai(t) being greater than or equal to 0.01. This facilitates the identification and calculation of the overall combination parameters in the early stages of the calculation, while in the later stages, the optimal data is sought from the better data, balancing the convergence speed and the search of all combinations.

[0069] Step 6: Convergence check

[0070] If the convergence condition is not met, return to step 3 and continue iterating; if the convergence condition is met, proceed to step 7.

[0071] If the global optimum J (gbest) fluctuates by less than 2% over 20 consecutive iterations, or the number of iterations reaches T=300.

[0072] Step 7: Output the optimal solution.

[0073] Output the power parameter combination corresponding to the global optimal position gbest, such as (580kW, 70%, 1100rpm, 2°), as an auxiliary scheme for ship power control.

[0074] Verification of the adaptability of the ship power optimization scenario of this invention: Taking the near-shore voyage optimization of a hybrid power vessel (diesel engine + battery) as an example, the algorithm application effect is as follows: Optimization results: voyage fuel consumption decreased by 3% (from 25t to 24.25t), and NOx emissions decreased by 7% (from 120kg to 111.6kg). Key adaptation points for applying the method of this invention: Facing changes in sea state (such as wind speed increasing from 5m / s to 10m / s), the power parameters are quickly fine-tuned (diesel engine power increased from 1000kW to 1500kW) to adapt to increased wind resistance; when the battery SOC approaches the 20% threshold, the charging power is finely adjusted (from 50kW to 120kW) to avoid deep discharge.

[0075] The method of this invention is applicable to scenarios such as voyage power distribution and real-time power control of ships with multiple power sources (diesel, battery, wind power, etc.). This invention can also be used in extreme and sudden operating conditions (such as power equipment failure), but it needs to be combined with emergency control logic (such as rapid switching to backup power). In the event of a sudden operating condition, the method of this invention is only used as an auxiliary optimization tool.

[0076] The method of this invention boasts high prediction accuracy: the prediction accuracy for power demand in the next 0.3 hours exceeds 92%, a significant improvement over traditional models (75% accuracy). It can predict changes in power demand 60-120 seconds in advance, allowing ample time for optimization decisions. The optimization effect of this invention is significant: the power scheme of this invention can reduce fuel consumption per voyage by 1%-4%, reduce emissions by 5%-8%, and avoid lifespan degradation caused by frequent equipment start-ups and shutdowns. The method of this invention is highly adaptable: through multi-source data fusion and online updates, the model can adapt to different navigation environments such as inland waterways and coastal waters, as well as different power combinations such as batteries and diesel, providing a universal optimization solution for various types of hybrid power vessels.

Claims

1. A method for optimizing ship dynamics, characterized in that: Includes the following steps: Step 1: Data Acquisition and Parameter Initialization Steps The data for the variables requiring optimization are standardized and parameters are set using the following formula: xinorm represents normalized data. i represents the original data. The maximum value in the data. The minimum value in the data; Step 2: Set the constraint range for the optimization variables: Based on the different ship performance characteristics, optimize the value range of different parameters. Step 3: Fitness Calculation For each particle, the J value is calculated using the multi-objective function, and the formula is as follows: minJ = ωy × A + ωy × B + ... + ωy × C, where y takes values ​​from 1 and 2 to the particle dimension; A, B, and C are the factor values, respectively. Step 4: Speed ​​and Position Update The new velocity vi(t+1) of each particle is calculated according to the velocity update formula; the formula is as follows: vi(t+1) = Ai(t)vi(t) + c1fi(xi(t)−gbest) + c2fi(xi(t)−pibest) vi(t) is the velocity vector of the i-th particle at time t; c1 and c2 are acceleration factors, fi is the pulse frequency of the i-th particle, fi = fmin + (fmax - fmin) × rand(), fmin and fmax are the upper and lower limits of the frequency, rand() is a random number in (0, 1); xi(t) is the position vector of the i-th particle at time t, gbes is the global optimal position vector of all particles, and pibest is the individual optimal position vector of the i-th particle. Calculate the new position: xi(t+1) = xi(t) + vi(t+1). If the new position exceeds the constraint range, then trim to the boundary. Step 5: Pulse parameter update: The updated formula calculates Ai(t+1) = α × Ai(t), where α is the attenuation coefficient. Step 6: Convergence check If the convergence condition is not met, return to step 3 and continue iterating; if the convergence condition is met, proceed to step 7. Step 7: Output the optimal solution. Output the combination of dynamic parameters corresponding to the globally optimal position gbest.

2. The method for optimizing ship dynamics according to claim 1, characterized in that: The optimization variables are diesel generator power, battery, blade angle of attack, and propeller speed.

3. The method for optimizing ship dynamics according to claim 2, characterized in that: The fitness calculation process is as follows: minJ=ω1×fuel+ω2×emission+ω3×speed+ω4×wind_efficiency, where ω1+ω2+ω3+ω4=1; Fuel is the power factor value of the diesel generator set; Emission is the battery life factor value; speed is the propeller speed factor value; and wind_efficiency is the wind energy utilization factor value.

4. A method for optimizing ship dynamics according to claim 1 or 2, characterized in that: The parameters are set as follows: population size = 700, maximum number of iterations T = 300, fmin = 1, fmax = 5, initial loudness Ai(0) = 1.

0.

5. The method for optimizing ship dynamics according to claim 2, characterized in that: The power output of the diesel generator set ranges from 80kW to 2000kW.

6. The method for optimizing ship dynamics according to claim 2, characterized in that: The battery SOC ranges from 20% to 90%.

7. The method for optimizing ship dynamics according to claim 2, characterized in that: The thruster speed ranges from 200 rpm to 1800 rpm.

8. The method for optimizing ship dynamics according to claim 2, characterized in that: The angle of attack of the blade ranges from -15° to 30°.

9. A method for optimizing ship dynamics according to claim 1 or 2, characterized in that: c1 = 2−2t / T, c2 = 2t / T, where t is the current iteration number and T is the maximum iteration number.

10. A method for optimizing ship dynamics according to claim 1 or 2, characterized in that: α = 0.95, Ai(t) is greater than or equal to 0.01.

Images