KR20230041519A - System and method for operating optimal posture of vessel on real sea area by using machine learning and computer-readable recording medium including the same - Google Patents

Abstract

Disclosed is a system for operating the optimal posture of a ship in a real-sea area using machine learning, comprising: a learning data DB (110) that collects and stores learning data on a ship posture for ship specifications, operating conditions, and marine environmental conditions; a machine learning unit (120) modeled with a machine learning algorithm for deriving the optimal posture of the ship for each ship specification, operating conditions during actual operation, and marine environmental conditions by machine learning the learning data in the learning data DB (110); a monitoring unit (130) that monitors the operating conditions and marine environmental conditions during actual operation of the ship in real time; and a control panel unit (150) that applies, to a ballast water control device (140), the optimal posture derived in real time according to the operating conditions and marine environmental conditions from the monitoring unit (130) during actual operations through the machine learning unit (120). Therefore, navigation efficiency can be improved by optimizing the posture of the ship.

Description

System and method for operating optimal attitude of ship in real sea area using machine learning, and computer readable recording medium on which program for executing the method is recorded on computer LEARNING AND COMPUTER-READABLE RECORDING MEDIUM INCLUDING THE SAME}

The present invention predicts and applies the optimal hull flow characteristics according to various operating conditions and maritime environmental conditions of each ship through machine learning of the flow characteristics around the hull according to various ship specifications, various operating conditions, and maritime environmental conditions, and operates the ship. The present invention relates to a system and method for operating an optimal attitude of a ship in a real sea area using machine learning, which can improve efficiency, and a computer-readable recording medium on which a program for executing the method in a computer is recorded.

In general, when a ship dynamically shakes back and forth, it is referred to as trim, and when the ship is operated in a state tilted to one side due to an imbalance between the stern and the bow of the ship, the operation efficiency is significantly reduced.

For example, if the front and rear balance is not balanced due to trim during sailing, the cargo may be tilted in one direction, which acts as a risk factor to cargo stability as well as operational efficiency. Accordingly, by manually measuring the inclination of the bow and stern, the pump and valve of the ballast water control device are operated to move the ballast water in the ballast tank to maintain the front-rear balance of the stern and bow of the ship.

On the other hand, referring to FIG. 1 as related prior art, Korean Patent Laid-open Publication No. 10-2015-0021970 is disclosed. According to the interaction between the fuel consumption, efficient energy and fuel consumption and energy production are to be optimized through the model (4) that simulates the function of ship performance and energy optimization.

However, it is limited to realizing efficient use of energy by simulating meteorological conditions, environmental conditions, and operating conditions without considering the characteristics of the flow of the vessel. In addition, in the conventional case, the optimum condition of the vessel through a model test in a towing tank Although the trim is estimated, there is a limit to estimating the optimal navigation attitude, especially the optimal trim, because the flow characteristics are different between the real scale and the model scale.

Accordingly, a technique capable of estimating an optimal trim under various conditions is required.

Korean Patent Laid-open Publication No. 10-2010-0056829 (Vessel with storage tank and ballast tank) Korean Patent Publication No. 10-2015-0021970 (Method and system for predicting ship performance) Korean Patent Registration No. 10-1535888 (System and method for active and passive stabilization of ships)

The technical problem to be achieved by the spirit of the present invention is to optimize the flow of the hull according to the various operating conditions and maritime environmental conditions of each ship through machine learning of the flow characteristics around the hull according to various ship specifications, various operating conditions, and maritime environmental conditions. Provision of a system and method for operating an optimal attitude of a ship in real sea area using machine learning, which can improve navigation efficiency by predicting and applying characteristics, and a computer-readable recording medium on which a program for executing the method is recorded on a computer there is

In order to achieve the above object, an embodiment of the present invention, learning data DB for collecting and storing the learning data of the ship attitude for the ship specifications and operating conditions and maritime environmental conditions; A machine learning unit, modeled with a machine learning algorithm, which machine learns the learning data of the collected learning data DB and derives the optimal attitude of the ship for each ship specification, operating conditions during actual operation and marine environment conditions; a monitoring unit for real-time monitoring of operating conditions and maritime environmental conditions during the actual operation of the ship; And controlling, through the machine learning unit, applying the optimal attitude of the ship derived in real time from the monitoring unit to the ballast water control device in real time according to the operating conditions and marine environment conditions during the actual operation from the monitoring unit. Provides a real sea area ship optimal attitude operating system using machine learning, including; panel unit.

In addition, the ship specifications are draft, trim, list, rudder angle, main engine horsepower, main engine rotational speed, shaft rotational speed, propeller thrust, auxiliary engine horsepower, and any one or more of the driving force of the driving unit of the ballast water control device Including, the navigation condition and the marine environment condition are provided from the navigation information integration system, the navigation condition includes any one or more of ship speed, turn angle, draft during operation, and rudder angle during operation, and the marine environment condition may include any one or more of wind speed, wind direction, wave height, wave speed, current, temperature, humidity, and vessel location.

In addition, the machine learning algorithm, simple / multiple linear regression, decision tree, Bayesian network, support vector machine, KNN, decision tree, random forest, artificial neural network, dimensionality reduction algorithm, gradient boosting algorithm, adaboost, perceptron and It can be performed by any one of deep learning.

In addition, the learning data may be collected through any one or a combination of simulation, model ship experiment, actual operation, and estimation formula.

In addition, the machine learning unit may be installed on land and transmit the optimum attitude information derived by the machine learning unit to the control panel unit of the ship through a communication device.

In addition, the machine learning unit may self-learn and upgrade the machine learning algorithm with the profile of the optimal attitude for the navigation conditions and marine environment conditions acquired through the actual navigation.

In addition, the learning data DB collects and stores learning data of the propulsion of the main engine of the ship and cargo stability for the ship specifications, the operating conditions, and the marine environment conditions, and the machine learning unit collects and stores the learning data of the operating conditions during the actual operation. It can be modeled with the machine learning algorithm that derives the propulsion force and cargo stability of the main engine providing the optimal attitude of the ship for the marine environment conditions.

In addition, the ballast water control device is composed of a ballast tank, the driving unit for supplying ballast water to the ballast tank, and a control unit for controlling the driving unit, and the control unit, through the machine learning unit, controls the actual operation Among them, the ship attitude of the trim and list may be optimized by controlling the driving unit to correspond to the optimal attitude derived in real time according to the operating conditions and marine environment conditions from the monitoring unit.

On the other hand, another embodiment of the present invention, collecting the learning data of the optimal attitude of the ship for the ship specifications, operating conditions and maritime environment conditions, and storing it in a learning data DB; Building a machine learning model with a machine learning algorithm that machine-learns the learning data of the collected learning data DB and derives an optimal attitude of the ship for each ship specification, operating conditions during actual operation and marine environment conditions; real-time monitoring of operating conditions and maritime environmental conditions during the actual operation of a ship equipped with a ballast water control device; deriving the optimal attitude corresponding to the trim or list of the ship in real time according to the monitored operating conditions and marine environment conditions during the actual navigation through the machine learning algorithm; and controlling the ballast water controller by applying the derived optimal attitude to the control unit in real time.

In addition, the ship specifications are draft, trim, list, rudder angle, main engine horsepower, main engine rotational speed, shaft rotational speed, propeller thrust, auxiliary engine horsepower, and any one or more of the driving force of the driving unit of the ballast water control device Including, the navigation condition and the marine environment condition are provided from the navigation information integration system, the navigation condition includes any one or more of ship speed, turn angle, draft during operation and rudder angle during operation, and the marine environment condition may include any one or more of wind speed, wind direction, wave height, wave speed, current, temperature, humidity, and vessel location.

In addition, the machine learning algorithm, simple / multiple linear regression, decision tree, Bayesian network, support vector machine, KNN, decision tree, random forest, artificial neural network, dimensionality reduction algorithm, gradient boosting algorithm, adaboost, perceptron and It can be performed by any one of deep learning.

In addition, the learning data may be collected through any one or a combination of simulation, model ship experiment, actual operation, and estimation formula.

In addition, the machine learning algorithm can be self-learned and upgraded with the profile of the optimal attitude for the navigation conditions and marine environment conditions acquired through the actual navigation.

In addition, the learning data DB collects and stores learning data of the propulsion of the main engine of the ship or cargo stability for the ship specifications, the operating conditions, and the marine environment conditions, and the machine learning unit collects and stores the operating conditions during the actual operation. It can be modeled with a machine learning algorithm that derives the propulsion or cargo stability of the main engine that provides the optimal attitude of the ship for the marine environment conditions.

On the other hand, another embodiment of the present invention provides a computer-readable recording medium on which a program for executing the method for operating the ship's optimal attitude in a real sea area using the machine learning described above is recorded in a computer.

According to the present invention, the flow characteristics around the hull according to various ship specifications, various operating conditions, and maritime environmental conditions are predicted and applied to the optimal hull flow characteristics according to various operating conditions and maritime environmental conditions of each ship through machine learning. It has the effect of improving flight efficiency.

In addition, by controlling the ballast water control device in rapidly changing sea conditions, it is possible to improve fuel efficiency and operational efficiency by realizing the optimal ship attitude, and to respond to newly built ships, new routes, and global environmental changes. There is an effect that the machine learning model can be continuously upgraded by self-learning the machine learning algorithm with the optimal ship attitude profile for the operational conditions and marine environmental conditions obtained through actual operation in the maritime climate change.

Furthermore, it is possible to optimally operate the ballast water control system even in new or predicted conditions and environments to realize optimal navigation performance.

1 illustrates a system for predicting the performance of a ship according to the prior art.
2 is a schematic configuration diagram of a system for operating an optimal attitude of a ship in a real sea area using machine learning according to an embodiment of the present invention.
FIG. 3 illustrates a ship to which the system for operating an optimal attitude of a ship in a real sea area using machine learning of FIG. 2 is applied.
4 is a schematic flowchart of a method for operating an optimal attitude of a ship in a real sea area using machine learning according to another embodiment of the present invention.

Hereinafter, embodiments of the present invention having the above-described characteristics will be described in more detail with reference to the accompanying drawings.

According to an embodiment of the present invention, the system for operating the optimum vessel posture in a real sea area using machine learning includes a learning data DB 110 for collecting and storing learning data of a vessel posture for ship specifications, operating conditions, and marine environment conditions; A machine learning unit 120, which is modeled with a machine learning algorithm that machine learns the learning data of the learning data DB 110 and derives the optimal attitude of the ship for each ship specification, operating conditions and maritime environment conditions during actual operation, Through the monitoring unit 130 and the machine learning unit 120 that monitor the operating conditions and marine environment conditions during actual operation of the ship in real time, during actual operation, according to the operating conditions and marine environment conditions from the monitoring unit 130 Including the control panel unit 150, which applies the optimal attitude of the ship derived in real time to the ballast water control device 140 in real time, the main point is to improve the navigation efficiency by optimizing the attitude of the ship.

Hereinafter, with reference to FIG. 2, a detailed description of a system for operating an optimal ship attitude in a real sea area using machine learning having the above-described configuration is as follows.

First of all, the learning data DB 110 is a data storage device for machine learning provided in the bridge, that is, the bridge, and machine learning for ship specifications, navigation conditions, and marine environment conditions corresponding to the input values of the machine learning unit 120. Ship attitude corresponding to the output value of unit 120, in particular, training data of trim is collected and stored.

Specifically, ship specifications include draft, trim (difference in locking level between bow and stern due to pitch motion), list (port and starboard bias due to roll motion), rudder angle, main engine horsepower, and main engine rotation number, shaft rotation speed, propeller thrust, auxiliary engine horsepower, and the driving force of the driving unit 142 of the ballast water control device 140 may include any one or more.

In addition, navigation conditions and maritime environment conditions may be provided from the navigation information integration system 160 provided in the ship, and the navigation conditions include ship speed and heading angle (HDG) by GPS speed or speed log. And includes any one or more of the draft or rudder angle during navigation, and the marine environment conditions may include any one or more of wind speed, wind direction, wave height, wave speed, current, temperature, humidity, and vessel location.

In addition, the above-mentioned learning data may be collected through any one or a combination of navigation computer simulation, model ship experiment in the tank, actual operation, and statistical estimation formula.

Next, the machine learning unit 120 is a computer for machine learning provided in the ship, particularly in the bridge, and performs machine learning on the learning data of the learning data DB 110 to determine the navigation conditions and marine environment conditions during actual operation by ship specification. It can be composed of a machine learning module modeled with a machine learning algorithm that derives an optimal ship attitude (optimal attitude), especially an optimal trim.

In addition, the machine learning unit 120 is not installed only on board, but may be installed on land. When the machine learning unit 120 is installed on land, the optimal posture information derived by the machine learning unit 120 is transmitted to the communication device. It may be transmitted to the control panel unit 150 of the ship through (not shown).

On the other hand, machine learning algorithms include simple/multiple linear regression, decision trees, Bayesian networks, support vector machines, KNN (K-Nearst Neighbors), and decision trees. Tree), Random Forest, Artificial Neural Network, Dimensional Reduction Algorithm, Gradient Boosting Algorithm, Adaboost, Perceptron and Deep Learning ( Deep learning).

In addition, the machine learning unit 120, in addition to the past data of the learning data DB 110, the navigation conditions and information obtained through actual navigation on a newly built ship, a new route, or maritime climate change according to global environmental changes. Optimal operation of the ballast water controller 140 in any new or predicted conditions and environments by self-learning the machine learning algorithm with the profile of the ship's optimal attitude for the marine environment conditions and continuously upgrading the machine learning model can also be implemented.

Meanwhile, navigation conditions may be collected from navigation information from ECDIS (Electronic Chart Display and Information System) and navigation warning information from CAMS (Central Alert Management System).

In addition, the learning data DB 110 collects and stores learning data of the driving force of the main engine or cargo stability for ship specifications, operating conditions, and marine environment conditions, and the machine learning unit 120 collects and stores operating conditions and marine environment during actual operation. In addition to demonstrating the optimal performance of the ballast water control device 140 through machine learning, modeled with a machine learning algorithm derived to predict the propulsion or cargo stability of the main engine that provides the optimal ship speed for environmental conditions, It can also be used for engine or cargo management.

Here, the cargo stability may be evaluated by stress distribution or cargo stability of the container due to a change in tensile force of the cargo securing device according to the type and number of securing devices (not shown).

Next, the monitoring unit 130 monitors real-time navigation conditions and marine environment conditions during actual navigation of the ship on which the ballast water control device 140 is disposed, and provides them as input values to the machine learning unit 120 .

Here, as illustrated in FIG. 3, the maritime environmental conditions of wind direction and speed may be measured and monitored in real time by the wind direction anemometer 131 installed on the bow, and information on wave height, wave speed, current, temperature or humidity may be measured by the corresponding instrument Alternatively, it can be measured and monitored in real time through the corresponding sensor, and other marine environmental conditions are forecast weather data from the European Meteorological Agency (ECMWF) or the server of each country adjacent to the flight route through communication equipment, or by preceding the same flight route. It can also be transmitted through meteorological information of actual meteorological data of wind direction/speed/atmospheric pressure/swell of another sailing ship.

Next, the ballast water control device 140 operates the bow and stern ballast pumps for the purpose of anti-healing, and automatically supplies the water in the ballast tank on the inclined side to the ballast tank on the other side by an amount corresponding to the degree of inclination. The difference in level of locking between the bow and the stern due to the pitch motion of the

For example, as illustrated in (a) of FIG. 3, the ballast water control device 140 controls ballast water to ballast tanks 141a and 141b and ballast tanks 141a and 141b disposed at the bow and stern, respectively. It consists of a drive unit 142 which is a ballast pump that supplies and a control unit 143 that controls the drive unit 142, and the control unit 143, through the machine learning unit 120, monitors the unit 130 during actual operation. ) by controlling the drive unit 142 to correspond to the optimal attitude derived in real time according to the operating conditions and marine environmental conditions from ), so that when the stern trim exceeds the reference value, ballast water is additionally supplied to the bow ballast tank 141a or Ballast water is discharged from the stern ballast tank (141b), and if the bow trim exceeds the standard value, ballast water is discharged from the bow ballast tank (141a) or ballast water is supplied to the stern ballast tank (141b), so that trim or list By optimizing the attitude of the ship, it is possible to enable stable and efficient navigation.

Referring to the schematic details of the ballast water control device 140 with reference to FIG. 3 (b), the ballast water control device 140 includes the bow and stern ballast tanks 141a and 141b, the ballast pipeline 144 , a bow and stern water level detection sensor (not shown), a ballast pump 142, and several valves 145, so that the ship is at an angle or more than a predetermined angle to either side of the bow or stern during sailing or unloading or loading work in the port In case of inclination, the degree is automatically detected by the water level detection sensor, and ballast water is replenished into the bow and stern ballast tanks 141a and 141b through the sea chest 146 and the corresponding valve 145 to or to discharge ballast water from the bow and stern ballast tanks 141a and 141b.

Next, the control panel unit 150 is installed on the bridge, and through the machine learning unit 120, during actual operation, the optimal attitude derived from the monitoring unit 130 according to the operating conditions and marine environment conditions of the ship in real time. It is applied to the ballast water control device 140 in real time to optimize the optimal attitude, in particular, the trim by the controller 143, so as to improve operational stability.

On the other hand, the optimal attitude of the ship may be different depending on the marine environment conditions in which the ship operates. That is, the optimal attitude of a ship according to marine environmental conditions such as the sea in a static state and the optimal attitude of a ship according to marine environmental conditions in a state of high wind direction, wind speed, wave height, and wave speed may be different. At this time, in the case of a ship operating in maritime environmental conditions with high wind direction, wind speed, wave height, and wave speed, dynamic stability is also an important factor as important as operational efficiency. In the case of marine environmental conditions with high wave speed, etc., it is desirable to derive the optimal attitude of the ship by considering (weighting) not only the operational efficiency but also the dynamic behavior of the ship (heave, pitch, roll, etc.).

4 is a schematic flowchart of a method for operating an optimal ship attitude in a real sea area using machine learning according to another embodiment of the present invention. Briefly, the details are as follows. Here, the operating method is a sequence for operating the aforementioned operating system, and descriptions overlapping with one embodiment will be omitted below.

First, in the learning data DB storage step (S110), the ship's optimal attitude for the ship's specifications, operating conditions, and marine environment conditions, in particular, the learning data of the optimal trim are collected and stored in the learning data DB 120.

Specifically, the ship specifications are draft, trim, list, rudder angle, main engine horsepower, main engine rotational speed, shaft rotational speed, propeller thrust, auxiliary engine horsepower, and of the driving unit 142 of the ballast water control device 140 Any one or more of the driving forces may be included.

In addition, the navigation conditions and marine environment conditions may be provided from the navigation information integration system 160 provided in the ship, and the navigation conditions are any one of the ship speed, turn angle, and draft or rudder angle during navigation by GPS speed or speed log. It includes one or more, and the marine environment conditions may include any one or more of wind speed, wind direction, wave height, wave speed, current, temperature, humidity, and vessel location.

In addition, the above-mentioned learning data may be collected through any one or a combination of navigation computer simulation, model ship experiment in the tank, actual operation, and statistical estimation formula.

Then, in the machine learning model construction step (S120), the learning data of the learning data DB 110 is machine-learned to provide the optimal attitude of the ship for each ship specification, operating conditions and maritime environment conditions during actual operation. A machine learning model is built with a machine learning algorithm that derives the optimal posture.

For example, machine learning algorithms include simple/multiple linear regression, decision trees, Bayesian networks, support vector machines, KNNs, decision trees, random forests, artificial neural networks, dimensionality reduction algorithms, gradient boosting algorithms, adaboosts, and perceptrons, It can be performed by any one of deep learning.

Thereafter, in the monitoring step (S130), operational conditions and marine environmental conditions during actual navigation of the ship equipped with the ballast water control device 140 are monitored in real time.

Thereafter, in the optimal attitude derivation step (S140), an optimal attitude corresponding to the ship's trim or list is derived in real time according to the monitored operating conditions and marine environment conditions during actual navigation through a machine learning algorithm.

Thereafter, in the real-time application step (S150), the controller 143 applies the optimal attitude derived above in real time to control the ballast water control device 140 to maintain the optimal attitude of the ship to improve navigation efficiency. .

In addition, by self-learning and upgrading the machine learning algorithm with the profile of the optimal attitude for the navigation conditions and marine environment conditions obtained through actual operation, the machine learning algorithm is used to more accurately and useful ship specifications, navigation conditions, and marine environment conditions. It is possible to derive the ship attitude for

In addition, the learning data DB 110 collects and stores learning data of the driving force of the main engine or cargo stability for ship specifications, operating conditions, and marine environment conditions, and the machine learning unit 120 collects and stores operating conditions and marine environment during actual operation. It can be modeled with a machine learning algorithm that derives the thrust of the main engine or cargo stability that provides the ship's optimal attitude for environmental conditions.

On the other hand, another embodiment of the present invention provides a computer-readable recording medium on which a program for executing the method for operating the ship's optimal attitude in a real sea area using the machine learning described above is recorded in a computer.

Therefore, by the configuration of the embodiments as described above, through machine learning of the flow characteristics around the hull according to various ship specifications, various operating conditions and maritime environmental conditions, optimal Improves operational efficiency by predicting and applying hull flow characteristics, controls the ballast water control device in rapidly changing sea conditions to realize optimal ship attitude, improves fuel efficiency and improves operational efficiency, and improves newly built By self-learning the machine learning algorithm with the optimal attitude profile of the vessel for the operational conditions and maritime environmental conditions obtained through actual operation on the maritime climate change according to the vessel, new sea route, and global environmental change, the machine learning model is continuously By upgrading to , the ballast water control system can be operated optimally even in new or predicted conditions and environments to realize optimal navigation performance.

The embodiments described in this specification and the configurations shown in the drawings are only one of the most preferred embodiments of the present invention, and do not represent all of the technical spirit of the present invention, so various equivalents that can replace them at the time of this application. It should be understood that there may be waters and variations.

110: learning data DB 120: machine learning part
130: monitoring unit 131: wind direction anemometer
140: ballast water control device
141a, 141b: bow and stern ballast tanks
142: drive unit 143: control unit
144: pipeline 145: valve
146: sea chest 150: control panel unit
160: flight information integration system S110: learning data DB storage step
S120: Machine learning model construction step S130: Monitoring step
S140: Optimal posture derivation step S150: Real-time application step

Claims (15)

a learning data DB for collecting and storing ship attitude learning data for ship specifications, operating conditions, and maritime environmental conditions;
A machine learning unit, modeled with a machine learning algorithm, which machine learns the learning data of the collected learning data DB and derives the optimal attitude of the ship for each ship specification, operating conditions and marine environment conditions during actual operation;
a monitoring unit for real-time monitoring of operating conditions and maritime environmental conditions during the actual operation of the ship; and
Through the machine learning unit, during actual operation, a control panel for applying the optimal attitude of the ship derived in real time according to the operating conditions and marine environment conditions during the actual operation from the monitoring unit to the ballast water control device in real time. Including;
Optimal attitude operation system for ships in real sea area using machine learning. According to claim 1,
The vessel specifications include one or more of draft, trim, list, rudder angle, main engine horsepower, main engine rotational speed, shaft rotational speed, propeller thrust, auxiliary engine horsepower, and driving force of the driving unit of the ballast water control device, ,
The navigation condition and the marine environment condition are provided from the navigation information integration system, and the navigation condition includes one or more of ship speed, turn angle, draft during operation, and rudder angle during operation, and the marine environment condition includes wind speed, Characterized in that it includes any one or more of wind direction, wave height, wave speed, current, temperature, humidity and vessel location,
Optimal attitude operation system for ships in real sea area using machine learning. According to claim 1,
The machine learning algorithm is simple/multiple linear regression, decision tree, Bayesian network, support vector machine, KNN, decision tree, random forest, artificial neural network, dimension reduction algorithm, gradient boosting algorithm, adaboost, perceptron and deep learning Characterized in that it is performed by any one of
Optimal attitude operation system for ships in real sea area using machine learning. According to claim 1,
Characterized in that the learning data is collected through any one or a combination of simulation, model ship experiment, actual operation, and estimation formula,
Optimal attitude operation system for ships in real sea area using machine learning. According to claim 1,
Characterized in that the machine learning unit is installed on land and transmits the optimal attitude information derived by the machine learning unit to the control panel unit of the ship through a communication device.
Optimal attitude operation system for ships in real sea area using machine learning. According to claim 1,
Characterized in that the machine learning unit self-learns and upgrades the machine learning algorithm with the profile of the optimal attitude for the navigation conditions and marine environment conditions obtained through the actual operation,
Optimal attitude operation system for ships in real sea area using machine learning. According to claim 1,
The learning data DB collects and stores learning data of the propulsion force and cargo stability of the main engine of the ship for the ship specifications, the operating conditions, and the marine environment conditions,
Characterized in that the machine learning unit is modeled with the machine learning algorithm for deriving the propulsion force and cargo stability of the main engine that provides the optimal attitude of the ship for operating conditions and marine environmental conditions during the actual operation,
Optimal attitude operation system for ships in real sea area using machine learning. According to claim 2,
The ballast water control device is composed of a ballast tank, the drive unit for supplying ballast water to the ballast tank, and a control unit for controlling the drive unit,
The controller, through the machine learning unit, controls the drive unit to correspond to the optimal attitude derived in real time according to the navigation conditions and marine environment conditions from the monitoring unit during the actual navigation to trim and list the ship posture. characterized by optimizing
Optimal attitude operation system for ships in real sea area using machine learning. Collecting learning data of the ship's optimal attitude for ship specifications, operating conditions, and marine environment conditions and storing it in a learning data DB;
Building a machine learning model with a machine learning algorithm that machine-learns the learning data of the collected learning data DB and derives an optimal attitude of the ship for each ship specification, operating conditions during actual operation and marine environment conditions;
real-time monitoring of operating conditions and maritime environmental conditions during the actual operation of a ship equipped with a ballast water control device;
deriving the optimal attitude corresponding to the trim or list of the ship in real time according to the monitored operating conditions and marine environment conditions during the actual navigation through the machine learning algorithm; and
Controlling the ballast water control device by applying the derived optimal posture to the control unit in real time; including,
A method for operating the optimal posture of a vessel in real sea area using machine learning. According to claim 9,
The vessel specifications include one or more of draft, trim, list, rudder angle, main engine horsepower, main engine rotational speed, shaft rotational speed, propeller thrust, auxiliary engine horsepower, and driving force of the driving unit of the ballast water control device, ,
The navigation condition and the marine environment condition are provided from the navigation information integration system, and the navigation condition includes one or more of ship speed, turn angle, draft during operation, and rudder angle during operation, and the marine environment condition includes wind speed, Characterized in that it includes any one or more of wind direction, wave height, wave speed, current, temperature, humidity and vessel location,
A method for operating the optimal posture of a vessel in real sea area using machine learning. According to claim 9,
The machine learning algorithm is simple/multiple linear regression, decision tree, Bayesian network, support vector machine, KNN, decision tree, random forest, artificial neural network, dimension reduction algorithm, gradient boosting algorithm, adaboost, perceptron and deep learning Characterized in that it is performed by any one of
A method for operating the optimal posture of a vessel in real sea area using machine learning. According to claim 9,
Characterized in that the learning data is collected through any one or a combination of simulation, model ship experiment, actual operation, and estimation formula,
A method for operating the optimal posture of a vessel in real sea area using machine learning. According to claim 9,
Characterized in that the machine learning algorithm is self-learned and upgraded with the profile of the optimal attitude for the operating conditions and marine environmental conditions obtained through the actual operation.
A method for operating the optimal posture of a vessel in real sea area using machine learning. According to claim 9,
The learning data DB collects and stores learning data of the propulsion of the main engine of the ship or cargo stability for the ship specifications, the operating conditions, and the marine environment conditions,
Characterized in that the machine learning unit is modeled with a machine learning algorithm for deriving the propulsion force or cargo stability of the main engine that provides the optimal attitude of the ship for the operating conditions and marine environmental conditions during the actual operation,
A method for operating the optimal posture of a vessel in real sea area using machine learning. A computer-readable recording medium on which a program for executing the method for operating the ship's optimum posture in a real sea area using machine learning according to any one of claims 9 to 14 is recorded in a computer.

Images