KR102539035B1 - Operational search and rescue method for coastal and maritime safety accidents using numerical simulation - Google Patents

Abstract

The present invention relates to a continuous operational search and rescue method for coastal and maritime safety accidents using numerical simulation. The continuous operational search and rescue method comprises: a maritime area zoning step; a numerical simulation performance step; an accident location specification step; a movement path prediction step; and a search and rescue step. According to the present invention, a search and rescue range and location can be quickly and accurately predicted.

Description

Continuous operation search and rescue method for coastal and maritime safety accidents using numerical simulation {OPERATIONAL SEARCH AND RESCUE METHOD FOR COASTAL AND MARITIME SAFETY ACCIDENTS USING NUMERICAL SIMULATION}

The present invention relates to a permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation. More specifically, assuming virtual safety accidents at regular time intervals, Artificial intelligence to quickly and accurately predict the search and rescue range and location based on the numerical simulation results in the event of a shipwreck or a drowning body due to actual coastal and maritime safety accidents by conducting particle tracking numerical simulations for each area defined in advance This study relates to a permanent search and rescue method for coastal and maritime safety accidents using numerical simulations.

Recently, as the number of leisure populations and ships on the coast and sea has rapidly increased, the risk of safety accidents along the coast and at sea has also increased.

In particular, the coastal water is shallow, the surrounding environment is complex, and there are many factors that threaten safety, so the possibility of safety accidents is relatively high.

Moreover, the increase in floating debris dumped on the coast becomes a factor that deteriorates the coastal environment and lowers the safety of marine activities.

In addition, small vessels operating in coastal waters often share the same sea route, and there are many cases where there are fishing workshops near the sea route. will increase

2 is a map showing the location and topography of coastal and maritime safety accidents on the Korean Peninsula from 2017 to 2020.

On the other hand, in preparation for distress or location tracking that occurs on the coast and at sea, the installation of a location tracking system for search and rescue is increasing.

However, people in distress on the coast and at sea are still not protected from death due to hypothermia and other accidents or disasters because they do not receive prompt rescue measures.

As a prior art for solving these problems, Korean Patent Registration No. 10-2186519 (2020.12.03. Notice) includes an equipment database unit for storing current status information on various types of search and rescue equipment; a search and rescue equipment determination unit that searches the equipment database and determines the most suitable search and rescue equipment for a specific accident area where a maritime accident has occurred among the search and rescue equipment; a search and rescue route determining unit for determining an optimal search and rescue route for the search and rescue equipment determined by the search and rescue equipment determining unit based on the environmental information of the specific accident area; and a search and rescue route monitoring unit that monitors the actual search and rescue route of the search and rescue equipment to determine whether the actual search and rescue route deviate from the optimal search and rescue route. has been

Search and rescue work depends on the water temperature, but for successful search and rescue work, it is necessary to arrive at the scene of the accident within one hour and complete the search and rescue work within six hours under normal circumstances to ensure survival.

However, the above prior art does not reflect the actual coastal and maritime environment, such as the flow pattern of ocean currents, in the search route when search target factors such as persons in distress and underfloor due to coastal and maritime safety accidents occur, and patterned By using the search route, there was a problem that the so-called golden time of rescue could be missed due to the decrease in speed and accuracy of the search and rescue range and location of the search target factor based on the location where the coastal and maritime safety accident occurred.

The present invention has been devised to solve the problems of the prior art as described above, and an object of the present invention is to virtualize a coastal activity site and sea area based on prior scenarios at regular time intervals (eg, 4 or 6 times a day). Coastal using numerical simulation that can quickly and accurately predict the search and rescue range and location of search target factors based on the results of numerical simulations conducted prior to the occurrence of actual coastal and maritime safety accidents by conducting particle tracking numerical simulations assuming safety accidents And to provide a permanent operation search and rescue method for maritime safety accidents.

The present invention has been devised to solve the problems of the prior art as described above, and the always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to the present invention can cause accidents including coastal accident-prone areas or coastal waters. a maritime area division step of dividing the sea into a plurality of areas according to a predetermined criterion; A numerical simulation progress step of assuming a virtual safety accident for each area divided in the marine area division step at a predetermined time interval and performing particle tracking numerical simulation based on a flow field reflecting an actual coastal and marine environment; an accident location specifying step of specifying a location where a safety accident occurs when a safety accident occurs in the accident-probable sea; Based on the position predicted in the accident location specification step, any one area among a plurality of areas divided in the marine area division step is specified, and the specified area is searched through particle tracking numerical simulation conducted in the numerical simulation progress step. a movement path prediction step of predicting an expected movement path of a target factor; a search and rescue step of searching for an element to be searched along the estimated movement route predicted in the movement route prediction step; It is characterized by including.

Here, the numerical simulation progress step is to apply the Lagrangian method that considers advection and diffusion based on the flow field updated at regular time intervals to the Particle Tracking Model (PTM) to perform particle tracking numerical simulation. characterized by

In addition, the permanent search and rescue method for coastal and maritime safety accidents using numerical simulation according to the present invention matches the numerical simulation result of the numerical simulation progress step with the actual location of the search target factor searched through the search and rescue step A numerical model optimization step of learning artificial intelligence and optimizing a particle tracking model used in the numerical simulation progress step; It is characterized by further including.

In addition, the permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation according to the present invention includes a ship location acquisition step of acquiring and updating ship location information of ships operating in the accident-probable sea in real time or periodically. ; Further comprising, in the accident location specification step, if a safety accident occurs in a ship operating on the accident-probable sea, the location of the safety accident based on the latest updated ship location information in the ship location acquisition step is determined. It is characterized by predicting and specifying.

Here, the vessel location acquisition step is characterized by using vessel location information periodically transmitted by a vessel to a maritime traffic control center (VTS) using maritime communication including V-Pass or AIS.

Here, the accident location specification step is characterized in that artificial intelligence learns the actual movement path of the ship and predicts the expected movement route of the ship and the location where the safety accident occurred by time based on the learned movement route of the ship.

In addition, in the permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation according to the present invention, artificial intelligence learns the numerical simulation results in the numerical simulation progress step, and based on the learned numerical simulation results, the flow field A dangerous situation notification step of predicting the possibility of occurrence of each type of safety accident of a vessel according to a flow pattern and notifying a dangerous situation to a vessel operating in the accident-probable sea; It is characterized by further including.

According to the configuration described above, the always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to the present invention divides the accident-probable sea into a plurality of areas, and in each divided area, virtual Assuming a safety accident, a particle tracking numerical simulation is performed for the coast and sea area in advance, and when an actual safety accident occurs, the location of the accident is specified, and the search and rescue is performed by predicting the moving path of the agent to be searched based on the result of the numerical simulation that has already been performed. By proceeding, it has the advantage of being able to quickly and accurately predict the search and rescue range and location in case of coastal and maritime safety accidents.

In addition, the always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to the present invention uses an atypical detailed numerical model in consideration of artificial intelligence to define a very detailed grid for the area of concern for accident occurrence, and In offshore areas where there are not many of them, the variable grid network divided into relatively large grids is mixed to improve the accuracy of the search and rescue range and location for the search target factor, and to analyze the flow pattern of the flow field before coastal and maritime safety accidents occur. It has the advantage of being able to notify dangerous situations in advance.

1 is a flowchart of a permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation according to an embodiment of the present invention.
Figure 2 is a map showing the location and topography of coastal and maritime safety accidents on the Korean Peninsula from 2017 to 2020
3 is an exemplary view of numerical simulation for each area in the numerical simulation progress step according to an embodiment of the present invention.
Figure 4 is a probability distribution range of particles over time in particle tracking numerical simulation
5 is a conceptual diagram showing a state of optimizing a particle tracking model through artificial intelligence learning in a numerical model optimization step according to an embodiment of the present invention.
6 is a flow chart of a permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation according to another embodiment of the present invention.
7 is a conceptual diagram showing a state in which an expected movement path of a ship is optimized through artificial intelligence learning in an accident location specification step according to another embodiment of the present invention.
8 is a graph showing a method of predicting a moving path of a ship using long-term and short-term memory
9 is a graph showing a method of predicting a moving path of a ship based on acceleration prediction

Hereinafter, the always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to the present invention will be described in more detail with reference to the embodiments shown in the drawings.

1 is a flowchart of an always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to an embodiment of the present invention.

Referring to FIG. 1, the always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to an embodiment of the present invention includes a maritime area division step (S10), a numerical simulation progress step (S20), and an accident location It includes a specific step (S50), a moving path prediction step (S60), a search and rescue step (S70), and a numerical model optimization step (S80).

The sea area division step (S10) is a step of dividing the accident-probable sea into a plurality of areas based on a predetermined standard.

Here, the accident-probable sea is a comprehensive concept including coastal accident-prone areas or coastal waters, and is not limited to the sea.

2 shows a map showing the location of coastal and maritime safety accidents on the Korean Peninsula from 2017 to 2020 and their topography.

As shown in FIG. 3, a certain criterion for dividing the accident-probable sea into a plurality of areas is to divide the latitude and longitude of the coastal waters in a horizontal * vertical format, divide it into specific zones according to a certain rule, or Areas can be divided into regions.

In addition, it is possible to define a very detailed grid for an accident-probable area, and use a variable grid network divided into a relatively large grid in an offshore area where there is not much vessel traffic.

In the numerical simulation progress step (S20), a virtual safety accident is assumed for each area divided in the marine area division step (S10) at regular time intervals, and particle tracking numerical simulation is performed based on the flow field reflecting the actual coastal and marine environment. It is a step in progress.

That is, the present invention is an active search and rescue method that assumes virtual safety accidents at regular time intervals as shown in FIG. is to present

Figure 4 shows the probability distribution range of particles over time in the particle tracking numerical simulation (source: Simulation of Marine Leisure Accidents Using Random-Walk Particle Tracking on Macro-Tidal Environment, thesis). It can be seen that the range of the probability distribution of the particles widens with the passage of .

That is, as shown in FIG. 4, it can be confirmed that the search and rescue range widens as time elapses, and the search and rescue probability decreases accordingly. An active search and rescue method is required.

In addition, in the case of a flow field reflecting actual coastal and marine environments, as an example, an ADCIRC+SWAN (finite element model) flow field that combines tidal-tide-wave updated at regular time intervals can be used, and various other marine numerical models are also available. can be utilized

The flow field as described above is actively updated in real time by identifying whether the artificial intelligence is updating the data for observation and numerical forecast data provided by the Korea Meteorological Administration, National Maritime Survey, JTWC (Joint Typhoon Warning Center), NOAA, etc. It is possible to implement a flow field reflecting the actual coastal and marine environment by acquiring LDAPS and GDAPS data provided by the Korea Meteorological Administration every 6 hours in real time.

In addition, in one embodiment of the present invention, the Lagrangian method considering advection and diffusion based on the flow field reflecting the actual coastal and marine environment is applied to the Particle Tracking Model (PTM) to prevent accidents in coastal and Under the premise that it may occur in the sea area, the particle tracking numerical simulation is conducted in a pre-divided area.

Matter floating on the ocean surface is treated as a free drift, i.e. neutral particle, with randomness depending on ocean currents and surface wind stress, regardless of shape and size.

Current random-walk particle tracking models (PTMs) are described by Bilgili et al. and Buyukcelebi et al., for identification of marine and estuary mixing mechanisms.

The present invention proceeds with particle tracking numerical simulation for each region by applying the Lagrangian method considering advection and diffusion to a particle tracking model (PTM) using an atypical detailed numerical model.

There are two methods for viewing the motion of a fluid: the Euler method and the Lagrangian method.

The Euler method is a method that focuses on a specific location in space where a fluid flows over time, the Lagrangian method is a method that follows the coordinates of an individual fluid's location as it moves through space and time, Euler focuses on space, Lagrangian method is a point of view that focuses on the substance (particle) itself.

Since the main point of the present invention is to use the results of particle tracking numerical simulation for search and rescue work, the numerical simulation is performed using the Lagrangian method, which focuses on particles.

Since the Lagrangian method, the particle tracking model (PTM), and the particle tracking numerical simulation themselves are known technologies in the art to which the present invention belongs, detailed descriptions thereof will be omitted.

Meanwhile, the particle tracking model (PTM) used in the numerical simulation progress step (S20) is optimized through a numerical model optimization step (S80) to be described later using artificial intelligence learning.

Currently, the National Institute of Meteorological Sciences of Korea is also developing technology to learn observation data and numerical model results using artificial intelligence in order to improve computation speed and secure reliability.

The accident location specifying step (S50) is a step of specifying a location where a safety accident occurs when a safety accident occurs in the accident-probable sea.

If the sea where an accident can occur is a coastal area, it is possible to specify the location where a safety accident occurred by collecting accident location information, such as a report from an accident witness or party.

On the other hand, when the sea where an accident can occur is a nearby sea, a location where a safety accident has occurred can be specified based on vessel location information, which will be described in more detail in the description of another embodiment of the present invention.

The movement path prediction step (S60) specifies any one area among a plurality of areas divided in the maritime area division (S10) step based on the location specified in the accident location specifying step (S50), and This is a step of predicting the expected movement path of the factor to be searched through the particle tracking numerical simulation performed in the numerical simulation progress step (S20) for the search object.

That is, in the movement path prediction step (S60), when an actual safety accident occurs at sea, the particle tracking numerical simulation result of the virtual safety accident previously performed in the flow field reflecting the actual sea situation at regular time intervals for the accident area is used. This is a step of quickly and accurately predicting the expected movement path of the factor to be searched.

The search and rescue step (S70) is a step of searching for an agent to be searched along the expected movement route predicted in the movement route prediction step (S60).

The conventional search and rescue method searches using a patterned search route through simple monitoring, and the speed and accuracy of the search and rescue range and location of the search target factor based on the location of coastal and maritime safety accidents are poor, so-called Although there was a problem that the golden time of the structure could be missed, the present invention can quickly and accurately search the search target element by searching the search target element along the expected movement path predicted in the movement path prediction step (S60). be able to

The numerical model optimization step (S80) introduces artificial intelligence to reflect the local characteristics of the tidal rise in tidal channels or gansaji, etc. to identify the coastal external force characteristics in a reliable manner, and as shown in FIG. 5, the numerical simulation progress step (S20 ) by matching the numerical simulation result (Forecast Data) with the actual location (Observation Data) of the search target factor searched through the search and rescue step (S70) to learn artificial intelligence, and to be used in the numerical simulation progress step (S20) This is the step of optimizing the particle tracking model.

The artificial intelligence learning part used in one embodiment of the present invention may use a long short term memory (LSTM) technique, which is a known technology.

The long-short-term memory is one of the recurrent neural network (RNN) techniques developed to prevent the vanishing gradient problem, which is a problem of existing recurrent neural networks (RNNs), by using cells, input gates, output gates, and forget gates.

In the above, the always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to an embodiment of the present invention has been reviewed, and hereinafter, another embodiment of the present invention will be reviewed.

6 is a flowchart of an always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to another embodiment of the present invention.

Referring to FIG. 6, the always-operated search and rescue method for coastal and maritime safety accidents using numerical simulation according to another embodiment of the present invention includes a maritime area division step (S10), a numerical simulation progress step (S20), and a vessel location Acquisition step (S30), dangerous situation notification step (S40), accident location specification step (S50), movement route prediction step (S60), search and rescue step (S70), and numerical model optimization step (S80) include

Another embodiment of the present invention is limited to the case where coastal and maritime safety accidents occur in ships, except for the ship location acquisition step (S30), the dangerous situation notification step (S40), and the accident location specification step (S50). Since is the same as an embodiment of the present invention, only parts different from the embodiment of the present invention will be described below.

The vessel location acquisition step (S30) is a step of acquiring and updating vessel location information of a vessel operating in the accident-probable sea in real time or periodically.

In another embodiment of the present invention, the vessel location acquisition step (S30) uses vessel location information periodically transmitted by a vessel to a maritime traffic control center (VTS) using maritime communication including V-Pass or AIS.

That is, this is to acquire and update vessel location information of ships operating in the accident-probable sea in real time or periodically by using ship location information that ships are obliged to transmit to the Maritime Control Center in accordance with the laws of the Republic of Korea.

In the dangerous situation notification step (S40), artificial intelligence learns the numerical simulation result in the numerical simulation progress step (S20), and based on the learned numerical simulation result, the possibility of occurrence by type of safety accident of the ship according to the flow pattern of the flow field It is a step of predicting and notifying a dangerous situation to a ship operating on the sea where an accident may occur.

As described above, the artificial intelligence learning part used in one embodiment of the present invention may use a long short term memory (LSTM) technique, which is a known technology.

Through the dangerous situation notification step (S40) as described above, the present invention prepares for dangerous situations by informing dangerous situations to ships operating on the sea where accidents can occur in advance before coastal and maritime safety accidents occur through flow pattern analysis of the flow field. By doing so, damage can be prevented in advance.

The accident location specification step (S50) is the location where the safety accident occurred based on the latest updated ship location information in the ship location acquisition step (S30) when a safety accident occurs in the ship operating the accident-probable sea. It is a step of predicting and specifying.

In one embodiment of the present invention, the accident location specification step (S50), as shown in FIG. 7, artificial intelligence learns the actual travel route of the ship, and based on the learned travel route of the ship, the ship It is configured to accurately predict and specify in real time the predicted travel route and the location where the safety accident occurred by hour.

As described above, the artificial intelligence learning part used in one embodiment of the present invention may use a long short term memory (LSTM) technique, which is a known technology.

In this regard, FIG. 8 is a graph showing a method for predicting a ship's movement path using long and short-term memory [Paper title: New ship movement path prediction method using long and short-term memory (Kim Jong-hee, Jeong Chan-ho, Kang Do-geun, Lee Chang-jin (2020) - Korean Institute of Electrical Engineers) )]am.

Here, LSTM-B on the left, LSTM-P on the right, black indicates the previous route, blue indicates the actual route, and red indicates the predicted route, and indicates examples of prediction results and prediction errors.

On the other hand, Figure 9 is a graph showing a method for predicting the movement path of a ship based on acceleration prediction [Paper title: Acceleration prediction-based new ship movement path prediction method (Kim Jong-hee, Jeong Chan-ho, Kang Do-geun, Lee Chang-jin (2020) - Korean Institute of Electrical and Electronics Engineers) ]am.

Here, each error is shown below each result, the left side is an LSTM-based method that directly predicts latitude and longitude, the middle is a method for predicting speed, which is the difference between latitude and longitude in the same three-layer short-term memory model, and the right side denotes an acceleration prediction-based method.

Through the accident location specification step (S50) as described above, it is possible to predict the search and rescue range and location with high accuracy by reducing the error with respect to the actual location in the event of a coastal or maritime safety accident.

The always-operated search and rescue method for coastal and maritime safety accidents using numerical simulations described above and shown in the drawings is only one embodiment for carrying out the present invention, and should not be construed as limiting the technical spirit of the present invention. Can not be done. The scope of protection of the present invention is determined only by the matters described in the claims below, and the improved and changed embodiments without departing from the gist of the present invention are obvious to those skilled in the art to which the present invention belongs. It will be said that it falls within the protection scope of the present invention.

S10 Maritime domain division step
S20 numerical simulation progress
S30 Vessel Location Acquisition Step
S40 Dangerous situation notification step
S50 Accident location specific step
S60 movement path prediction step
S70 Search and rescue phase
S80 numerical model optimization step

Claims (7)

A maritime area division step of dividing a coastal accident-prone area or an accident-probable sea including coastal waters into a plurality of areas according to a predetermined standard;
A numerical simulation progress step of assuming a virtual safety accident for each area divided in the marine area division step at a predetermined time interval and performing particle tracking numerical simulation based on a flow field reflecting an actual coastal and marine environment;
an accident location specifying step of specifying a location where a safety accident occurs when a safety accident occurs in the accident-probable sea;
Based on the location specified in the accident location specifying step, any one area among a plurality of areas divided in the marine area division step is specified, and the specified area is searched through particle tracking numerical simulation conducted in the numerical simulation progress step. a movement path prediction step of predicting an expected movement path of a target factor;
a search and rescue step of searching for an element to be searched along the estimated movement route predicted in the movement route prediction step; A permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation, characterized in that it comprises. According to claim 1,
In the numerical simulation step, the particle tracking numerical simulation is performed by applying the Lagrangian method that considers advection and diffusion based on the flow field updated at regular time intervals to the Particle Tracking Model (PTM). A permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulations. According to claim 2,
A numerical model optimization step of matching the numerical simulation result of the numerical simulation step with the actual position of the search target factor searched through the search and rescue step to learn artificial intelligence and to optimize the particle tracking model used in the numerical simulation step. cast; A permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation, characterized in that it further comprises. According to claim 1,
A vessel location acquisition step of acquiring and updating vessel location information of a vessel operating in the accident-probable sea in real time or periodically; include more,
The accident location specification step is to predict and specify the location where the safety accident occurred based on the latest updated ship location information in the ship location acquisition step when a safety accident occurs in the ship operating the accident-probable sea A permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation, characterized in that. According to claim 4,
The vessel location acquisition step uses vessel location information periodically transmitted by the vessel to the maritime traffic control center (VTS) using maritime communication including V-Pass or AIS, and coastal and maritime safety accidents using numerical simulation. An always-on search and rescue method for According to claim 4,
In the accident location specification step, artificial intelligence learns the actual movement route of the ship, and based on the learned movement route of the ship, predicts and specifies the expected movement route of the ship and the location where the safety accident occurred by time A permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation. According to any one of claims 4 to 6,
A vessel that navigates the accident-probable sea by predicting the possibility of occurrence by type of safety accident of the vessel according to the flow pattern of the flow field based on the artificial intelligence learning of the numerical simulation results in the numerical simulation progress stage and based on the learned numerical simulation results. A dangerous situation notification step that can notify a dangerous situation to; A permanent operation search and rescue method for coastal and maritime safety accidents using numerical simulation, characterized in that it further comprises.

Images