Bulk cargo ship multi-mode self-adaptive power propulsion control method and system

Abstract

The invention discloses a multi-mode self-adaptive power propulsion control method and system for a bulk cargo ship, relates to the technical field of power propulsion control of bulk cargo ships, and aims to solve the problems of poor power control adaptability and low energy efficiency of an existing bulk cargo ship. According to the method, ship navigation data are collected in real time through a multi-source sensor, and ship working condition features are generated through preprocessing and feature extraction; according to the ship working condition characteristics, the working mode where the ship should be located is judged by fusing a basic rule base and an intelligent reasoning mechanism; and calling a control strategy corresponding to the working mode of the ship to adjust the main engine and the adjustable propeller, and synchronously implementing safety monitoring on mode switching. Efficient and low-carbon operation of the power system of the bulk cargo ship is achieved, the operation safety and reliability of the bulk cargo ship are improved, and the requirements of different navigation working conditions are met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bulk carrier power propulsion control, in particular to a bulk carrier multi-mode adaptive power propulsion control method and system. BACKGROUND

[0002] As the core carrier for global transportation of bulk commodities such as coal, ore and grain, the operation efficiency, working condition adaptation ability and operation safety of the power propulsion system of the bulk carrier are directly related to the control of the ship operation cost and the compliance with environmental protection. At present, the bulk carrier power propulsion control in the industry generally adopts a segmented control strategy based on fixed parameters, that is, a limited power mode is determined according to a preset typical navigation scenario, and the main engine speed and propeller pitch angle are adjusted by a traditional PID controller to realize basic power output matching.

[0003] However, the existing control strategy has significant technical shortcomings and cannot adapt to the actual operation requirements of the bulk carrier. On the one hand, the working condition adaptation is seriously insufficient. The bulk carrier needs to face dynamic changes in load, environmental disturbances and navigation stage switching during navigation, and the power demand varies greatly. The traditional fixed mode cannot dynamically adjust the control logic according to real-time working condition parameters. On the other hand, the energy efficiency and environmental protection compliance are poor. There is a strong coupling relationship between the main engine thermal efficiency, propeller efficiency and working condition parameters. The traditional control only takes speed or sailing speed as a single control target and does not optimize the parameters in combination with the optimal running interval of the main engine thermal efficiency and propeller efficiency, resulting in high fuel consumption per unit distance. It cannot meet the mandatory requirements of the International Maritime Organization on fuel intensity index by 2025, and shipowners face the risk of high carbon tax or operation restrictions. At the same time, the control precision and safety are insufficient. In the boundary working condition, the fixed threshold switching logic is easy to produce mode misjudgment, causing frequent fluctuations in main engine power, and the rate of change of power parameters is not limited during mode switching, which easily causes damage to the shafting and further increases the safety hazards of navigation. Therefore, the existing control strategy cannot meet the efficient, low-carbon and reliable operation requirements of the bulk carrier, and an adaptive power propulsion control scheme is urgently needed. SUMMARY

[0004] The present application aims to provide a bulk carrier multi-mode adaptive power propulsion control method and system to solve the problems raised in the background.

[0005] To solve the above technical problems, the technical solution adopted by the present application is:

[0006] The bulk carrier multi-mode adaptive power propulsion control method comprises the following steps:

[0007] S1. Real-time acquisition of ship navigation data by deploying multi-source sensors on the ship, wherein the ship navigation data includes main engine operating parameters, propulsion system state parameters, ship motion parameters and environmental disturbance parameters;

[0008] S2. Preprocessing and feature extraction of ship navigation data to generate ship working condition features;

[0009] S3. Based on the ship working condition features, the working mode of the ship should be judged by fusing the basic rule base and the intelligent reasoning mechanism, including economic mode, high-speed mode, low-load mode and maneuvering mode;

[0010] S4. Calling the control strategy corresponding to the working mode of the ship to adjust the ship actuator, and synchronously implementing safety monitoring, the ship actuator including the main engine and the adjustable propeller.

[0011] Preferably, the main engine operating parameters include the main engine speed collected by the speed encoder, the main engine torque collected by the torque sensor and the main engine fuel consumption rate collected by the fuel flowmeter; the propulsion system state parameters include the propeller speed collected by the propeller speed sensor and the shaft power collected by the shaft power instrument; the ship motion parameters include the speed collected by the GPS positioning instrument, the draft collected by the draft sensor, the trim angle collected by the inertial measurement unit and the ship heading collected by the heading sensor; the environmental disturbance parameters include the wind direction collected by the wind direction sensor and the wave height collected by the wave height instrument.

[0012] Preferably, the preprocessing method is as follows:

[0013] Adaptive Kalman filtering is used for noise filtering of high-frequency fluctuation data, including main engine operating parameters, propulsion system state parameters and speed, and sliding median filtering with a window length of 5-10s is used for noise filtering of low-frequency drift data, including draft, trim angle, ship heading and environmental disturbance parameters. The high-frequency fluctuation data and low-frequency drift data after noise filtering are identified for outliers by the Isolation Forest algorithm, and the effective data of 3-4 continuous time points before and after the outliers are linearly interpolated to repair, to obtain the preprocessed ship navigation data.

[0014] Preferably, the ship working condition features include the main engine load rate , the main engine thermal efficiency , the propeller efficiency , the net load change rate , the trim change rate , the relative wind direction angle and the effective wave height ; wherein the main engine load rate is calculated by bringing the preprocessed main engine speed and the main engine torque into the main engine load rate formula, and the main engine thermal efficiency is calculated by bringing the preprocessed main engine torque host engine speed and fuel consumption rate The host engine thermal efficiency is calculated by bringing the pre-processed propeller speed, shaft power and speed The net load change rate is calculated by bringing the pre-processed draft The trim change rate is calculated by bringing the pre-processed trim angle into the trim change rate formula, and the relative wind direction angle is calculated by bringing the pre-processed wind direction and ship heading into the relative wind direction angle formula, the effective wave height is calculated by bringing the pre-processed wave height into the wave height formula;

[0015] The host engine load rate formula is: wherein, is the rated power of the host engine;

[0016] The host engine thermal efficiency formula is: = wherein, is the low heat value of the fuel;

[0017] The propeller efficiency formula is: wherein, is the ship thrust, which is derived from the pre-processed propeller speed, shaft power and hydrodynamic model;

[0018] The net load change rate formula is: wherein, is the time interval, is the draft load weight conversion coefficient, is the light draft;

[0019] The trim change rate formula is: wherein, and are the pre-processed trim angles at the front and rear moments;

[0020] The relative wind direction angle formula is: wherein, when > , take ;

[0021] The wave height formula is: wherein, is the root mean square value of the pre-processed wave height data.

[0022] Preferably, the method for determining the working mode of the ship by fusing the basic rule base with the intelligent inference mechanism is:

[0023] S31. calling the basic rule base to match the working mode of the ship with the working condition characteristics, wherein the basic rule base pre-stores the matching threshold corresponding to each working mode, the matching threshold is the preset trigger condition initial parameter range of each working mode, and the trigger condition initial parameter corresponds to the working condition characteristics of the ship;

[0024] S32. If the matching threshold of a working mode covers the working condition characteristics of the ship, outputting the working mode as the working mode of the ship, otherwise, starting fuzzy logic inference, and selecting the working mode with the highest matching probability from each working mode as the working mode of the ship by defining the low, medium and high fuzzy subsets of the working condition characteristics of the ship and the triangular membership function;

[0025] S33. introducing a reinforcement learning algorithm, and constructing a reward function with unit fuel consumption, wherein the value of the reward function increases as the unit fuel consumption decreases, and accordingly, the reinforcement learning agent iteratively optimizes the matching threshold of each working mode in the basic rule base according to the preset sailing period, and the preset sailing period is 100-200km.

[0026] Preferably, the preset trigger condition initial parameter range of each working mode is:

[0027] The preset trigger condition initial parameter range of the economic mode is: main engine load rate 50%-80%, main engine thermal efficiency ≥ 38%, relative wind direction angle < 45°, and effective wave height ≤ 2m;

[0028] The preset trigger condition initial parameter range of the high-speed mode is: main engine load rate 80%-95%, main engine thermal efficiency ≥ 40%, and effective wave height ≤ 4m;

[0029] The preset trigger condition initial parameter range of the low load mode is: main engine load rate < 50%, propeller efficiency < 65%, and net load change rate ≤ 0.5t / h;

[0030] The preset trigger condition initial parameter range of the maneuvering mode is: main engine load rate 40%-70%, and trim change rate > 0.3 , and relative wind direction angle ≤

[0031] Preferably, the control strategy corresponding to the working mode of the ship is specifically:

[0032] ​​The control strategy corresponding to the economic mode is: a model predictive control algorithm based on a host thermodynamic model is used for the host, the minimum fuel consumption rate of the host is taken as the optimization objective, the host load rate is adjusted to [65%, 70%] by dynamically adjusting the fuel injection timing and the intake air amount, and the pitch angle of the adjustable propeller is adjusted to [ , ] by solving the thrust torque matching equation through iteration to obtain the optimal matching point of the thrust torque, i.e., the optimal target pitch angle.

[0033] The control strategy corresponding to the low load mode is: a PID control with speed closed loop is used, the fuel injection amount is adjusted by feeding back the deviation between the real host speed and the target host speed in the low load mode, the minimum stable speed of the host is adjusted to [40%, 50%] of the rated speed of the host, and the pitch angle of the adjustable propeller is adjusted to [ , ] to match the low power output of the host.

[0034] The control strategy corresponding to the high speed mode is: a PID feedforward composite control is used for the host, the feedforward link compensates the fuel supply amount in advance according to the speed deviation, the PID link suppresses the host speed fluctuation, the host speed overshoot is controlled within 5%, the pitch angle of the adjustable propeller is adjusted to [ , ] by using the PID control with shaft power feedback, and the pitch angle change rate of the adjustable propeller is controlled within 2 by output limiting.

[0035] The control strategy corresponding to the maneuvering mode is: a PID control is used for the host, the host speed adjustment delay is shortened by increasing the proportional coefficient of the PID controller, the host speed adjustment response time is controlled within 5s, and a nonlinear sliding mode control is used for the adjustable propeller to control the pitch angle tracking error of the adjustable propeller within .

[0036] Preferably, the process of synchronously implementing safety monitoring is:

[0037] When the working mode is switched, an S-shaped speed command curve is generated by a transition process controller, the transition process controller is based on the host power, propeller thrust of the current working mode and the preset host rated power ratio and propeller thrust of the target working mode, the host power increment per second is calculated through a preset smooth transition algorithm, and an S-shaped speed command curve is formed according to the host power increment per second, so that the host power change rate is limited within 10% of the host rated power per second; at the same time, the propeller thrust fluctuation before and after the working mode switching is detected through the thrust sensor, if the propeller thrust fluctuation is greater than 5% of the rated propeller thrust, the original working mode is returned and the sound and light alarm of the driver's cabin is triggered.

[0038] Due to the adoption of the above technical scheme, the technical progress achieved by the present application relative to the prior art is:

[0039] 1. The present application significantly improves the working condition adaptive matching capability of the bulk cargo ship power propulsion system, solves the problem that the existing fixed segmented control cannot adapt to complex navigation working conditions. The existing technology relies on preset fixed power mode of typical scenes, and it is difficult to cope with the differences in power demand brought by changes in load, environmental disturbances and navigation stage switching during navigation of the bulk cargo ship; and the present application fuses the basic rule base and intelligent reasoning mechanism, first realizes fast matching of conventional working conditions by using pre-stored matching thresholds of each working mode, then processes the boundary fuzzy working conditions through fuzzy logic reasoning, and iteratively optimizes the matching thresholds regularly with the unit fuel consumption as the target, so that the control logic can dynamically adapt to different navigation scenes, and reduce the problems of energy waste or insufficient response.

[0040] 2. The present application greatly improves the operating energy efficiency and environmental compliance of the bulk cargo ship power propulsion system, breaking through the limitations of the existing technology which only takes speed or speed as a single control target and does not couple the efficiency optimal interval. The existing technology does not combine the efficiency optimal interval adjustment of the host and the propeller, resulting in high unit fuel consumption, which is difficult to meet the requirements of the International Maritime Organization carbon intensity regulation; and the present application designs special high-efficiency control strategies for four types of working modes, helping the bulk cargo ship to meet environmental compliance and reduce the operating cost of the bulk cargo ship.

[0041] 3. The present application effectively improves the safety and reliability of the bulk cargo ship power propulsion control, solves the problems of boundary condition misjudgment, mode switching impact and lack of safety protection in the prior art. The existing technology is prone to misjudge the mode in the boundary condition due to the fixed threshold, and the power parameter change rate is not limited during switching, which may cause shaft impact damage; the present application generates an S-shaped speed command curve through a transition process controller when the mode is switched, controls the smooth change of the host power, and monitors the thrust fluctuation through a sensor, and if the stable threshold is exceeded, the original mode is returned and an alarm is triggered; in addition, the precision of each mode control is strengthened, which can avoid the risk of operation and improve the safety of navigation. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0043] Figure 1 The method flowchart of the present application. DETAILED DESCRIPTION

[0044] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0045] Embodiments, such as Figure 1 The bulk cargo ship multi-mode adaptive power propulsion control method comprises the following steps:

[0046] S1. Real-time collection of ship navigation data through multi-source sensors deployed on the ship, wherein the ship navigation data comprises main engine operating parameters, propulsion system state parameters, ship motion parameters and environmental disturbance parameters;

[0047] S2. Preprocessing and feature extraction of the ship navigation data to generate ship working condition features;

[0048] S3. Based on the ship working condition features, the working mode of the ship should be determined through the fusion of the basic rule base and the intelligent reasoning mechanism, wherein the working mode comprises an economic mode, a high-speed mode, a low-load mode and a maneuvering mode;

[0049] S4. Calling the control strategy corresponding to the working mode of the ship to adjust the ship actuator, and synchronously implementing safety monitoring, wherein the ship actuator comprises a main engine and an adjustable propeller.

[0050] Further, the working principle of the present application will be illustrated by the following embodiments:

[0051] In this embodiment, a 70,000-ton bulk cargo ship is taken as a test object. The rated power of the bulk cargo ship is 12,000 kW, the rated rotating speed is 100 r / min, and the rated thrust is 850 kN. The bulk cargo ship is used for coal transportation between a port in China and a port in Australia. The whole route of the coal transportation is about 6,800 nautical miles, and needs to go through open sea, narrow channel and port water area, and faces load change (full load draft 11.8 m and ballast draft 5.2 m) and environmental disturbance (wind direction and speed change, current direction and speed change, etc.). and wave height 0.5m-4.5m), the system hardware used contains a multi-source sensing module, a data processing module, a mode decision module, an execution control module and a safety monitoring module; among them, the multi-source sensing module contains 12 types of sensors such as a speed encoder, a torque sensor, a fuel flow meter, etc., which are respectively deployed at the main engine crankshaft, the fuel pipeline, the top of the bridge, etc.; the data processing module contains Advantech UNO-2484G embedded controller, and pre-installs MATLAB R2023a and Python library; the mode decision module integrates MySQL basic rule library, fuzzy reasoning and reinforcement learning algorithm; the execution control module contains main engine controller MAN ME Control System and propeller hydraulic servo controller; the safety monitoring module contains STM32H743 transition process controller and sound-light alarm.

[0052] The multi-source sensors deployed on the ship collect ship navigation data in real time at a sampling frequency of 10Hz, among them, the main engine running parameters are collected by the speed encoder on the free end of the main engine crankshaft, such as the main engine speed at a certain moment is 85r / min, the strain torque sensor on the output shaft of the main engine collects the main engine torque, such as the main engine torque at a certain moment is 32000N・m, the mass fuel flow meter on the fuel supply pipeline collects the main engine fuel consumption rate, such as the main engine fuel consumption rate at a certain moment is 185kg / h; the propeller system state parameters are collected by the magneto-electric speed sensor on the shaft end of the propeller, such as the propeller speed at a certain moment is 42.5r / min, and the propeller transmission ratio is 2:1, the shaft power instrument on the middle of the propeller shaft collects the shaft power, such as the shaft power at a certain moment is 8200kW; the ship motion parameters are collected by the GPS positioning instrument on the top of the bridge, such as the ship speed at a certain moment is 7.8m / s, the mean value of the draft is collected by the static pressure type draft sensor on the bottom of the bow, the middle and the stern, such as the draft at a certain moment is 10.8m, the trim angle is collected by the inertial measurement unit at the center of gravity of the ship, such as the trim angle at a certain moment is , the heading is collected by the fiber-optic gyro heading sensor of the bridge navigation station, such as the trim angle at a certain moment is ; the environmental disturbance parameters are collected by the ultrasonic wind direction sensor on the top of the mast, such as the wind direction at a certain moment is , the wave height is collected by the radar wave height instrument on the bow deck, such as the wave height at a certain moment is 1.5m, and the ship navigation data collected accordingly is transmitted in real time to the data processing module through 100Mbps ship Ethernet, and stored in 512GB SSD for subsequent processing.

[0053] The ship navigation data is divided into high-frequency fluctuation data with fluctuation frequency > 0.1 Hz and low-frequency drift data with fluctuation frequency ≤ 0.1 Hz, wherein the high-frequency fluctuation data includes main engine operating parameters, propulsion system state parameters and speed, and the low-frequency drift data includes draft, trim angle, ship heading and environmental disturbance parameters, the adaptive Kalman filter is used for the high-frequency fluctuation data, and the sliding median filter with a window length of 8s is used for the low-frequency drift data; subsequently, the isolated forest algorithm with 100 trees, a sample subset of 256 and an outlier ratio of 0.01 is used to identify outliers in the high-frequency fluctuation data and the low-frequency drift data after noise filtering, and the effective data of three continuous time points before and after the outliers are taken for linear interpolation repair; the ship working condition characteristics are calculated based on the pretreated ship navigation data:

[0054] The main engine load rate is calculated according to the main engine load rate formula (T = 32000 N·m, n = 85 r / min, = 12000 kW, and the result is approximately 72.2%), the main engine thermal efficiency is calculated according to the main engine thermal efficiency formula = (B = 185 kg / h, Q = 42000 kJ / kg, and the result is approximately 41.5%), the propeller efficiency is calculated according to the propeller efficiency formula (F ≈ 720 kN, v = 7.8 m / s, P = 8200 kW, and the result is approximately 68.5%), the net load change rate is calculated according to the net load change rate formula (k = 2500 t / m, d = 10.8 m, = 5.2 m, Δt = 1 h, and the result is 14000 t / h), the trim change rate is calculated according to the trim change rate formula ( = -0.5°, = -0.45°, Δt = 10 s, and the result is 0.005° / s), the relative wind direction angle is calculated according to the relative wind direction angle formula ( = 100°, = 125°, and the result is 25°), the effective wave height is calculated according to the wave height formula ( ≈ 2.2525 m², and the result is approximately 1.8 m), and finally the ship working condition characteristics are generated and transmitted to the mode decision module.

[0055] The basic rule base stored in the MySQL database in the calling mode decision module is called, which pre-stores the initial parameter ranges of the preset trigger conditions of the four working modes: the economic mode is that the host load rate is 50%-80%, the host thermal efficiency is ≥38%, the relative wind direction angle is <60°, and the effective wave height is ≤2m; the high-speed mode is that the host load rate is 80%-95%, the host thermal efficiency is ≥40%, and the effective wave height is ≤4m; the low load mode is that the host load rate is <50%, the propeller efficiency is <65%, and the net load change rate is ≤0.5t / h; the maneuvering mode is that the host load rate is 40%-70%, the trim change rate is >0.3° / s, and the relative wind direction angle is ≤90°. The generated ship working condition characteristics {72.2%, 41.5%, 68.5%, 14000t / h, 0.005° / s, 25°, 1.8m} are matched, all threshold values of the economic mode cover the characteristic set, and the economic mode is directly output as the working mode of the ship; if the basic rule base matching fails, such as the ship working condition characteristics at a certain moment are {79.5%, 39%, 2.1m,...}, the fuzzy logic reasoning based on the Scikit-Fuzzy library is started, the low, medium and high fuzzy subsets of the ship working condition characteristics and the triangular membership function are defined, the matching probabilities of each mode are calculated, such as the economic mode probability ≈0.516 and the high-speed mode ≈0.0268, and the one with the highest probability is selected; at the same time, the deep Q network reinforcement learning algorithm based on the Stable Baselines3 library is introduced, the reward function is constructed with R=100-0.01×(FC / V) (FC is the fuel consumption per unit time, and V is the speed), the basic rule base threshold values are iteratively optimized by the reinforcement learning agent according to the 150km preset sailing period, such as the host load rate range of the economic mode is optimized from 50%-80% to 55%-75%, so that the fuel consumption per unit distance is reduced by about 4%.

[0056] The control strategy corresponding to the economic mode is called to adjust the main engine and the adjustable propeller. The MPC (Model Predictive Control) based on Mean Value Engine Model (MVE model) is used for the main engine, the minimum fuel consumption rate is taken as the target, the prediction time domain is set to 5s, the control time domain is set to 2s, the fuel injection timing is set to ±2°CA, the intake amount is set to ±10% of the rated value, the optimal control amount is solved through the quadratic programming algorithm, the instruction is output to the MAN ME Control System every 0.1s, the fuel injection timing is adjusted from 12°CA to 10°CA, the intake amount is increased from 800kg / h to 850kg / h, the load rate of the main engine is reduced from 72.2% to 68%, falls into the interval [65%, 70%], the fuel consumption rate is reduced from 185kg / h to 172kg / h; the PSO (Particle Swarm Optimization) is used for the adjustable propeller, the maximum propeller efficiency is taken as the target, the particle number is set to 30, the iteration number is set to 20 and the inertia weight is set to 0.7, the optimal pitch angle 15.2° is obtained by solving the thrust-torque matching equation through iteration, the instruction is output to the hydraulic servo controller of the adjustable propeller to control the hydraulic cylinder to adjust the blade angle, the adjustment time is about 3s, the propeller efficiency is increased from 68.5% to 72%; the safety monitoring is implemented synchronously, it is assumed that the ship is switched from the economic mode to the high-speed mode at this time, the STM32H743 transition process controller generates the S-shaped speed instruction curve, receives the current main engine power 8200kW and the target power 10000kW of the high-speed mode, calculates the power increment 100kW / s per second, the power increment per second is ≤10% of the rated power of the main engine per second, and the power instruction is issued in 18 steps; the propeller thrust fluctuation amount before and after the working mode switching is monitored through the thrust sensor, if the propeller thrust fluctuation amount > 5% of the rated thrust, for example, the fluctuation amount is 5.4% this time, the economic mode is immediately returned and the audible and visual alarm of the bridge and the engine room is triggered.

[0057] The above is a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any modification or replacement within the technical range disclosed by the present application can be easily thought by those skilled in the art, which should be covered in the protection scope of the present application.

Claims

1. A multi-mode adaptive propulsion control method for bulk carriers, characterized in that, Includes the following steps: S1. Real-time ship navigation data is collected by multi-source sensors deployed on the ship, including main engine operating parameters, propulsion system status parameters, ship motion parameters, and environmental disturbance parameters; S2. Preprocess and extract features from ship navigation data to generate ship operating condition features; S3. Based on the characteristics of ship operating conditions, the ship's appropriate operating mode is determined by integrating a basic rule base and an intelligent reasoning mechanism. The operating modes include economic mode, high-speed mode, low-load mode, and maneuver mode. S4. Invoke the control strategy corresponding to the working mode of the ship to adjust the ship's actuators and implement safety monitoring simultaneously. The ship's actuators include the main engine and the adjustable propeller.

2. The multi-mode adaptive propulsion control method for bulk carriers according to claim 1, characterized in that, The main engine operating parameters include main engine speed, main engine torque, and main engine fuel consumption rate; the propulsion system status parameters include propeller speed and shaft power; the ship motion parameters include speed, draft, trim angle, and ship heading; and the environmental disturbance parameters include wind direction and wave height.

3. The multi-mode adaptive propulsion control method for bulk carriers according to claim 2, characterized in that, The preprocessing method is as follows: Adaptive Kalman filtering is used for noise filtering of high-frequency fluctuation data, and sliding median filtering with a window length of 5-10s is used for noise filtering of low-frequency drift data. Based on this, the noise-filtered high-frequency fluctuation data and low-frequency drift data are used to identify outliers through the isolated forest algorithm, and linear interpolation is performed on the effective data of 3-4 consecutive time points before and after the outlier to obtain preprocessed ship navigation data. The high-frequency fluctuation data includes main engine operating parameters, propulsion system status parameters, and speed, while the low-frequency drift data includes draft, trim angle, ship heading, and environmental disturbance parameters.

4. The multi-mode adaptive propulsion control method for bulk carriers according to claim 3, characterized in that, The ship's operating characteristics include main engine load rate, main engine thermal efficiency, propeller efficiency, net deadweight rate of change, trim rate of change, relative wind direction angle, and significant wave height. Specifically, the main engine load rate is calculated using pre-processed main engine speed and torque; the main engine thermal efficiency is calculated based on pre-processed main engine torque, main engine speed, and fuel consumption rate; the propeller efficiency is calculated based on pre-processed propeller speed, shaft power, and ship speed; the net deadweight rate of change is calculated based on the pre-processed draft change over time; the trim rate of change is the pre-processed trim angle of change over time; the relative wind direction angle is the angle between the pre-processed wind direction and the ship's heading; and the significant wave height is calculated based on the pre-processed wave height.

5. The multi-mode adaptive propulsion control method for bulk carriers according to claim 4, characterized in that, The method for determining the appropriate working mode of a ship by integrating a basic rule base and an intelligent reasoning mechanism is as follows: S31. Call the basic rule base to perform working mode matching on the ship's working condition characteristics. The basic rule base pre-stores the matching threshold corresponding to each working mode. The matching threshold is the preset trigger condition initial parameter range of each working mode. S32. If the matching threshold of a certain working mode covers the characteristics of the ship's working condition, then output the working mode as the working mode that the ship should be in; otherwise, start fuzzy logic reasoning to select the working mode with the highest matching probability from each working mode as the working mode that the ship should be in. S33. Introduce a reinforcement learning algorithm and construct a reward function based on fuel consumption per unit distance. The reinforcement learning agent periodically optimizes the matching threshold of each working mode in the basic rule base according to a preset flight cycle.

6. The multi-mode adaptive propulsion control method for bulk carriers according to claim 5, characterized in that, The initial parameter range for the preset trigger conditions of each working mode is as follows: The preset trigger conditions for the economic mode are as follows: host load rate 50%-80%, host thermal efficiency ≥38%, relative wind direction angle < And the effective wave height is ≤2m; The preset trigger conditions for high-speed mode are: host load rate 80%-95%, host thermal efficiency ≥40%, and effective wave height ≤4m. The preset trigger conditions for low-load mode are: main engine load rate < 50%, propeller efficiency < 65%, and net load change rate ≤ 0.5 t / h; The preset trigger conditions for the maneuver mode have the following initial parameter ranges: main engine load rate 40%-70%, and pitch change rate >0.

3. and relative wind angle ≤ .

7. The multi-mode adaptive propulsion control method for bulk carriers according to claim 6, characterized in that, The specific control strategy corresponding to the requested operating mode of the vessel is as follows: The control strategy corresponding to the economic mode is as follows: A model predictive control algorithm is used for the main engine, with the goal of minimizing the main engine fuel consumption rate, adjusting the main engine load rate to [65%, 70%]. Simultaneously, a particle swarm optimization algorithm is used for the adjustable propeller, with the goal of maximizing propeller efficiency, to solve for the optimal thrust-torque matching point in real time, thereby adjusting the pitch angle of the adjustable propeller to […]. , ]; The control strategy for low-load mode is as follows: PID control is used to adjust the minimum stable speed of the main engine to [40%, 50%] of the rated speed of the main engine, while simultaneously adjusting the pitch angle of the adjustable propeller to []. , ]; The control strategy corresponding to the high-speed mode is as follows: PID feedforward composite control is used for the main engine to keep the main engine speed overshoot within 5%, while PID control is used to adjust the pitch angle of the adjustable propeller to […]. , [and control the pitch angle variation rate of the adjustable propeller within 2] Within; The control strategy corresponding to the maneuvering mode is as follows: Increase the proportional gain of the PID controller for the main engine to control the main engine speed adjustment response time within 5 seconds; simultaneously, employ nonlinear sliding mode control for the adjustable propeller to control the propeller pitch angle tracking error within a specified range. Within.

8. The multi-mode adaptive propulsion control method for bulk carriers according to claim 7, characterized in that, The process of synchronously implementing security monitoring is as follows: When switching operating modes, an S-shaped speed command curve is generated through the transition process controller to limit the main engine power change rate to within 10% of the main engine rated power per second. At the same time, the propeller thrust fluctuation before and after the operating mode switch is detected and monitored. If the propeller thrust fluctuation is greater than 5% of the propeller rated thrust, the system will revert to the original operating mode and trigger an audible and visual alarm.

9. A multi-mode adaptive propulsion control system for bulk carriers, the system being used to implement the multi-mode adaptive propulsion control method for bulk carriers as described in claim 1, comprising: Multi-source sensing module, used to collect ship navigation data in real time by deploying multi-source sensors; The data processing module is used to preprocess and extract features from ship navigation data to generate ship operating condition features; The mode decision module is used to determine the appropriate working mode for a ship based on its operating condition characteristics by integrating a basic rule base with an intelligent reasoning mechanism. The execution control module is used to call the control strategy corresponding to the ship's operating mode to adjust the main engine and adjustable propeller; The safety monitoring module is used to monitor the safety of the ship's operating mode switching.

Images