METHOD AND SYSTEM FOR SUPPORTING THE MANAGEMENT OF A LIQUID GAS TRANSPORT VESSEL OF THE TYPE THAT CONSUMES EVAPORATED GAS FOR ITS PROPULSION

Description

Domaine technique

[0001] The invention relates to the field of liquefied gas transport vessels of the type that consume vaporized gas for propulsion. More specifically, the invention provides a method and a system for assisting in the management of such a vessel. Arrière-plan technologique

[0002] Leak-proof and thermally insulated tanks are commonly used for transporting Liquefied Natural Gas (LNG) at approximately -162°C at atmospheric pressure. These tanks can be used for transporting liquefied gases. Many other liquefied gases can also be considered, including methane, ethane, propane, butane, ammonia, hydrogen, and ethylene.

[0003] Ship tanks can be single or double membrane-sealed tanks that allow transport at atmospheric pressure. The membranes are generally made of thin sheet stainless steel or Invar. One membrane is usually in direct contact with the liquefied gas.

[0004] Because liquefied gas is transported at a temperature much lower than ambient temperature, it naturally tends to warm up slowly during transport on a ship carrying one or more such tanks, despite the tanks' thermal insulation. The gas generated as a result of this warming is commonly called boil-off gas (BOG). For example, during the transport of LNG at approximately -162°C at atmospheric pressure, BOG generation can be observed at a rate of approximately 0.05% to 0.15% per day of the initial volume of LNG contained in a tank.

[0005] To make use of BOG, it is known to utilize all or part of it for ship propulsion by supplying it to a ship's propulsion engine capable of consuming BOG. ​​Ships are also typically equipped with a Gas Combustion Unit (GCU) to burn any excess BOG that cannot be consumed by the ship's propulsion engine.

[0006] The fuel oil burned in the fuel cell during the ship's journey represents a loss, as it is not delivered to the ship's destination and is consumed without profit. Therefore, it is desirable to minimize the amount of fuel oil burned along a given ship's route. Furthermore, even if the fuel oil sent to the ship's propulsion engine is recovered, it also represents a reduction in the amount delivered to the ship's destination.

[0007] Documents FR3013672A1, EP3677498A1, US2019 / 241244A1 and US2015 / 324714A1 describe computer-implemented processes for assisting in the management of a liquefied gas transport vessel according to the prior art.

[0008] Also, there is a need for ship management assistance solutions that limit the total amount of BOG generated in the tank, so as to limit the amount of BOG to be burned in the GCU and the amount of BOG sent to the ship's propulsion engine. Résumé

[0009] One idea underlying the invention is to use weather forecasts along a predicted ship's route and a ship operating model to estimate at least a total quantity of evaporation gas generated in the ship's tank. Another idea underlying the invention is to perform this estimation for at least one, and preferably several, tank management scenarios defining a change in the pressure of the gas phase contained in the tank along said predicted route. Yet another idea underlying the invention is to display to a user at least one of said tank management scenarios as a function of a cost function that depends at least on a total quantity of evaporation gas generated in the tank along said predicted route.

[0010] According to one embodiment, the invention provides a computer-implemented method for assisting in the management of a liquefied gas transport vessel, the vessel comprising at least one tank configured to contain liquefied gas and a vapor phase treatment system, the vapor phase treatment system being capable of sending evaporated gas exiting the tank to a propulsion engine of the vessel and to a gas combustion unit on board the vessel, the vapor phase treatment system further being capable of extracting a portion of the liquid phase contained in the tank and evaporating this portion in order to send it to the propulsion engine, the method comprising: provide an initial state of the tank, said initial state of the tank including an initial pressure of the gaseous phase contained in the tank and an initial temperature of the liquid phase contained in the tank; provide a predicted route of the ship; from meteorological forecasts, determine at least one environmental parameter of the ship along said predicted route; determine a speed curve of the ship along said predicted route; generate a tank management scenario defining an evolution of the pressure of the gaseous phase contained in the tank along said predicted route, and from the tank management scenario thus generated: a) estimate a power curve required at the output of the propulsion engine along said predicted route as a function of said speed curve of the ship and said at least one environmental parameter;b) estimate a curve of the quantity of evaporated gas generated in the tank along said predicted path; c) estimate, from said power requirement curve and said curve of quantity of evaporated gas generated, a curve of the quantity of gas to be extracted from the tank and a curve of the quantity of evaporated gas to be burned in the gas combustion unit; and d) calculate a cost function that depends at least on a total quantity of evaporated gas generated in the tank along said predicted path; and display to a user the tank management scenario as a function of the calculated cost function.

[0011] Thanks to these features, a tank management scenario can be evaluated by calculating the cost function and displayed to the user in order to offer decision support.

[0012] Depending on the execution methods, such a process may include one or more of the following characteristics.

[0013] According to one execution mode, a plurality of tank management scenarios are generated, with steps a) to d) being performed for each tank management scenario, and at least one of said tank management scenarios is displayed to the user according to the cost function calculated for said tank management scenario.

[0014] Thus, a large number of tank management scenarios can be evaluated by calculating the cost function, and one or more tank management scenarios that best minimize the cost function can be displayed to the user in order to offer decision support.

[0015] According to one execution mode, before displaying at least one of said tank management scenarios to the user, said tank management scenarios are iteratively regenerated a plurality of times by a first evolutionary algorithm using the cost function as the first objective function.

[0016] An evolutionary algorithm (also known as an evolutionary algorithm) is a method, typically implemented by computer, in which a population of solutions is generated, each solution is evaluated using an objective function, some of the solutions that best minimize the objective function are selected, a new population of solutions is generated from these selected solutions, and these steps are repeated until a stopping criterion is met. In the context of the present invention, each tank management scenario is a solution, and the cost function serves as the objective function for the evolutionary algorithm.

[0017] Several types of evolutionary algorithms are known as such. In one execution mode, an evolutionary algorithm is a genetic algorithm. Evolutionary algorithms are particularly well-suited to optimization problems.

[0018] As just mentioned, an evolutionary algorithm necessarily uses a stopping criterion to decide when to stop searching for more optimal solutions. Depending on the execution mode, the stopping criterion is a computation time criterion, which ensures that the process will display one or more optimized tank management scenarios to the user after a given computation time.

[0019] According to one execution mode, the ship's predicted journey is divided into consecutive time intervals, and the tank scenario or scenarios define an average value for the pressure of the gaseous phase contained in the tank during each of said consecutive time intervals.

[0020] According to one execution mode, said at least one tank management scenario displayed to the user is shown together with a minimum value of the gas phase pressure contained in the tank and / or a maximum value of the gas phase pressure contained in the tank during each of said consecutive time intervals.

[0021] According to one execution method, said planned journey comprises journey stages defined each by two waypoints and a heading to be followed between said two waypoints, and at stage a), the power curve required at the output of the propulsion engine along said planned journey is estimated as a function of said headings, said ship speed curve and said at least one environmental parameter.

[0022] In this way, the evaluation of tank management scenarios also takes into account the impact of the ship's heading on the power required at the propulsion engine output. Indeed, for example, depending on the angle between the ship's heading and the direction of the current, wind, and waves, this power can be greater or lesser.

[0023] According to one execution method, said ship speed curve is determined from said predicted route and said at least one environmental parameter by a second evolutionary algorithm using a second objective function which depends on said total quantity of evaporation gas generated in the tank along said predicted route and a difference between a required time of arrival at destination and an estimated time of arrival at destination.

[0024] In this way, the estimation of the power required at the output of the propulsion engine takes into account a required time of arrival at the ship's destination, which can be specified as an input parameter by a user.

[0025] A liquefied gas unloading terminal systematically requires that the tank have a given unloading pressure and / or temperature. In one embodiment, providing a predicted vessel journey includes providing an unloading pressure and / or temperature required by an unloading terminal at the vessel's destination, and in step (d), the cost function further depends on a difference between the pressure of the gas phase contained in the tank at the end of said predicted journey and said unloading pressure and / or a difference between the temperature of the liquid phase contained in the tank at the end of said predicted journey and said unloading temperature.

[0026] In this way, a tank management scenario is considered all the more acceptable as it allows arrival at the unloading terminal located at the destination point with an unloading pressure and / or an unloading temperature close to those required by the unloading terminal.

[0027] According to one execution mode, in step a), the required power curve is estimated using a first statistical model, the first statistical model being capable of estimating a required power output of the propulsion engine at least as a function of a ship speed setpoint and said at least one environmental parameter, and the first statistical model being trained by a supervised machine learning method.

[0028] Supervised learning is a machine learning method that teaches a prediction function from annotated examples. In other words, a supervised learning method builds a predictive model based on a series of examples where the expected outcome is known. Supervised learning is typically implemented by computer; the initial training of the statistical model is also typically performed by computer.

[0029] It is understood that in the phrase or characteristic "train a statistical model by a supervised machine learning method on a test dataset", the "test dataset" may also include or consist of data from so-called "real" campaigns, i.e., data obtained or collected on ships operating as carriers and users of liquefied gas.

[0030] According to one embodiment, the aforementioned step b) of estimating a curve of quantity of evaporation gas generated in the tank along said predicted path includes estimating a first quantity of evaporation gas generated in the tank as a function of a current pressure of the gaseous phase contained in the tank and a current temperature of the liquid phase contained in the tank and estimating a second quantity of evaporation gas generated in the tank as a function of a current pressure of the gaseous phase contained in the tank and said at least one environmental parameter.

[0031] Thus, the estimation of the quantity of evaporation gas generated in the tank takes into account both the generation of evaporation gas linked to the coexistence of the liquid phase and the gaseous phase in the tank, and the generation of evaporation gas due to the agitation of the liquid phase due to the movements undergone by the ship.

[0032] According to one execution mode, the second quantity of evaporation gas generated in the tank is estimated using a second statistical model, the second statistical model being trained by a supervised machine learning method.

[0033] According to one embodiment, the initial state includes an initial composition of the liquefied gas contained in the tank, and step b) of estimating a curve of quantity of evaporation gas generated in the tank along said predicted path further includes estimating a composition of the evaporation gas generated in the tank.

[0034] A liquefied gas cargo is almost never chemically pure: it is generally a mixture of several liquefied gases, one of which is the predominant gas by mass, while the other(s) are present as impurities. However, the initial composition of the cargo is usually indicated by the liquefied gas supplier when it is loaded into the tank. By indicating the initial composition of the liquefied gas and estimating the composition of the vaporized gas generated in the tank, we take into account that the vaporized gas composition will change along the planned route due to the different vaporization properties of the gases composing the liquefied gas cargo.

[0035] According to one execution mode, the environmental parameter includes at least one of a current direction, a current speed, a wind speed, a wind direction, a mean wave height, a wave direction, and a wave period.

[0036] According to one embodiment, the vapor phase treatment system further comprises a liquefier capable of reliquefying evaporation gas exiting the tank and returning the evaporation gas thus reliquefied to the tank, and the aforementioned step c) comprises estimating, from said power requirement curve and said evaporation gas quantity curve generated, a curve of quantity of gas to be extracted from the tank, a curve of quantity of evaporation gas to be burned in the gas combustion unit, and a curve of quantity of evaporation gas to be sent to the liquefier.

[0037] According to one embodiment, the vapor phase treatment system further comprises a subcooler capable of subcooling a portion of the liquid phase contained in the tank and returning the subcooled portion to the tank, and wherein the aforementioned step c) comprises estimating, from said power requirement curve and said generated evaporation gas quantity curve, a gas quantity curve to be extracted from the tank, an evaporation gas quantity curve to be burned in the gas combustion unit, and a liquid phase quantity curve to be extracted from the tank and sent to the subcooler.

[0038] According to one method of execution, the steps of the process are repeated at fixed intervals and / or upon receipt of new weather forecasts.

[0039] Thus, the process regularly reassesses tank management scenarios as the vessel progresses along its planned route and new weather forecasts become available. As a complement or alternative, the process steps can be repeated at the user's request.

[0040] According to one embodiment, the invention further provides a management aid system for a liquefied gas transport vessel, the vessel comprising at least one tank configured to contain liquefied gas and a vapor phase treatment system, the vapor phase treatment system being capable of sending evaporated gas exiting the tank to a propulsion engine of the vessel and to a gas combustion unit on board the vessel, the vapor phase treatment system being further capable of extracting a portion of the liquid phase contained in the tank and evaporating this portion in order to send it to the propulsion engine, the management aid system comprising: an acquisition unit configured to acquire an initial state of the tank and a predicted course of the vessel, said initial state of the tank comprising an initial pressure of the gaseous phase contained in the tank and an initial temperature of the liquid phase contained in the tank; a first communication unit configured to obtain weather forecasts from a weather forecast provider; and a computing unit configured to: from the weather forecasts obtained by the first communication unit, determine at least one environmental parameter of the vessel along said predicted course; determine a velocity curve of the vessel along said predicted course;generate a tank management scenario defining an evolution of the pressure of the gas phase contained in the tank along said predicted route, and from the tank management scenario thus generated: a) estimate a power curve required at the output of the propulsion engine along said predicted route as a function of said ship speed curve and said at least one environmental parameter; b) estimate a curve of the quantity of evaporative gas generated in the tank along said predicted route; c) estimate, from said power curve and said curve of quantity of evaporative gas generated, a curve of the quantity of gas to be extracted from the tank and a curve of the quantity of evaporative gas to be burned in the gas combustion unit; d) calculate a cost function that depends on at least a total quantity of evaporative gas generated in the tank along said predicted route;and a display unit configured to show a user the tank management scenario based on the calculated cost function.

[0041] Such a system provides the same advantages as the process described above.

[0042] According to one execution mode, said planned journey comprises journey stages defined each by two waypoints and a heading to be followed between said two waypoints, and wherein the computing unit is configured to a) estimate a power curve required at the output of the propulsion engine along said planned journey as a function of said headings, said ship speed curve and said at least one environmental parameter.

[0043] According to one execution mode, the acquisition unit and the display unit are onboard the ship, the first communication unit and the computing unit are located in a shore station, and wherein the system further comprises a second communication unit configured to put the acquisition unit and the computing unit into communication. Brief description of the figures

[0044] The invention will be better understood, and other objects, details, features and advantages thereof will become more apparent from the following description of several particular embodiments of the invention, given solely by way of illustration and not limitation, with reference to the accompanying drawings. [ fig.1 ] There figure 1 is a schematic representation of a liquefied gas transport vessel. fig. 2 ] There figure 2 is a block diagram schematically representing the tanks of the ship of the figure 1 as well as a vapor phase treatment system, a ship propulsion engine and a ship-mounted gas combustion unit. fig. 3 ] There figure 3 is a block diagram of a ship management support system for the figure 1 . [ fig. 4 ] There figure 4 is a block diagram of a ship management support system for the figure 1 according to a variant of the figure 3 . [ fig. 5A ] There figure 5A is a block diagram illustrating the first steps of a ship management assistance process. figure 1 , the process can be implemented by the management support system of the figure 3 . [ fig. 5B ] There figure 5B is a block diagram illustrating the following steps in the process of the figure 5A . [ fig. 6 ] There figure 6 represents, as an illustration, an example of the ship's planned route. fig. 7 ] There figure 7 is a graph illustrating an example of a tank management scenario generated in the tank management assistance system figure 3 . [ fig. 8 ] There figure 8 is a block diagram representing, for explanatory purposes, the models used by the management support system of the figure 3 to evaluate a given tank management scenario. fig. 9 ] There figure 9 is a graph representing, as an illustration, an example of a tank management scenario similar to that of the figure 7 , to which minimum and maximum pressure values ​​in the tank are associated. fig. 10 ] There figure 10 is a graph representing, as an illustration, an example of a tank management scenario similar to that of the figure 7 , subdivided into sub-intervals of time. fig. 11 ] There figure 11 is a graph illustrating an example of a speed management scenario generated in the vehicle management assistance system figure 3 in certain execution modes. fig. 12 ] There figure 12 is a block diagram representing, for explanatory purposes, the models used by the management support system of the figure 3 to evaluate a given speed management scenario. fig. 13 ] There figure 13 is a block diagram analogous to that of the figure 2 representing a variant of the vapor phase treatment system. fig. 14 ] There figure 14 is a block diagram analogous to that of the figure 2 representing another variant of the vapor phase treatment system. fig. 15 ] There figure 15 is a block diagram analogous to that of the figure 2 representing yet another variant of the vapor phase treatment system. Description des modes d'exécution

[0045] The following embodiments are described in relation to a vessel comprising a double hull forming a load-bearing structure in which a plurality of watertight and thermally insulated tanks are arranged. In such a load-bearing structure, the tanks may, for example, have a polyhedral geometry, such as a prismatic shape.

[0046] Such sealed and thermally insulated tanks are designed, for example, for the transport of liquefied gas. Liquefied gas is stored and transported in such tanks at a low temperature, which necessitates thermally insulated tank walls to maintain the liquefied gas at that temperature.

[0047] Such watertight and thermally insulated tanks also include an insulating barrier anchored to the ship's double hull and carrying at least one watertight membrane. For example, such tanks can be manufactured using technologies marketed under the applicant's Mark III® or NO96® trademarks, or other similar technologies.

[0048] There figure 1 illustrates a vessel 1 comprising four sealed and thermally insulated tanks 2. The four tanks 2 may have identical or different filling states.

[0049] There figure 2 is a block diagram schematically representing the four tanks 3, 4, 5, 6 of the ship of the figure 1 As depicted on the figure 2 , ship 1 is further equipped with a vapor phase treatment system 10, an engine 40 and a gas combustion unit (GCU) 30.

[0050] Engine 40 provides the power required for the propulsion of vessel 1; alternatively, it can provide both the power required for the propulsion of vessel 1 and the power required to supply electrical power to other equipment on vessel 1 (commonly referred to as "Hotel Load"). Engine 40 is capable of consuming gas in the vapor phase, which may originate from the boil-off gas (BOG) exiting tanks 3, 4, 5, and 6, or from a heater 60 designed to vaporize liquefied gas extracted from the tanks. Such an engine 40 is known as such and is therefore not described in detail here.

[0051] As is known in itself, the GCU 30 is capable of burning any excess BOG that could not be consumed by the engine 40. Such a GCU 30 is known as such and is therefore not described in detail here.

[0052] The vapor phase treatment system 10 is capable of sending BOG exiting each tank 3, 4, 5, 6 to the engine 40 or to the GCU 30, as appropriate. To accomplish this, the vapor phase treatment system 10 comprises a vapor phase treatment subsystem 13, 14, 15, 16 for each of the tanks 3, 4, 5, 6, and a BOG distribution unit 20 communicating with these subsystems 13, 14, 15, 16 so as to allow the BOG to be sent to the engine 40 or the GCU 30, as appropriate. Such vapor phase treatment systems and subsystems are known as such and are therefore not described in detail here.

[0053] Each subsystem 13, 14, 15, 16 is also capable of extracting a portion of the liquid phase contained in the corresponding tank 3, 4, 5, 6 and sending this portion to the heater 60 in order to evaporate it when desired.

[0054] Although we have represented a ship 1 equipped with four tanks, it is understood that ship 1 may be equipped with a different number of tanks, in particular a single tank or any number of tanks.

[0055] In the following, for convenience, we will describe the management assistance process 300 by referring to tank 3, it being understood that the management assistance process 300 can be implemented simultaneously for several or each of the tanks 3, 4, 5, 6.

[0056] There figure 3 illustrates an example of a management support system 100 on board ship 1. This management support system 100 includes a central unit 110 connected to a communication interface 130, a human-machine interface 140 and a database 150.

[0057] The communication interface 130 allows the central unit 110 to communicate with remote devices, for example to obtain meteorological data, ship position data, or other data.

[0058] The human machine interface 140 includes a display means 41. The display means 41 allows one or more tank management scenarios 50 to be displayed to a user, as will be detailed later.

[0059] The human-machine interface 140 also includes an acquisition means 42 allowing the user to manually provide quantities to the central unit 110 as will be detailed later.

[0060] The management support system 100 also includes a database 150. This database 150 can be used to store models, including models 61, 62, 63 as will be detailed later.

[0061] There figure 4 This illustrates an example of a land-based management support system 200 communicating with vessel 1. Vessel 1 comprises a central processing unit 210, a human-machine interface 140, and a communication interface 230. The central processing unit 110, the communication interface 130, and the database 150 are located in a land-based station 1000. The operation of the management support system 200 is similar to that of the management support system 100, differing only in that the results of calculations performed by the land-based central processing unit 110 are sent to the human-machine interface 140 on board vessel 1 via the communication interfaces 130 and 230. For example, the communication interfaces 130 and 230 can use terrestrial or satellite radio frequency data transmission.

[0062] By referring to figures 5A à 9 We will now describe a 300 method for assisting in the management of ship 1, which can be implemented using the management assistance system 100 or 200.

[0063] The process 300 includes a first step 301 of providing an initial state of the tank 3. This initial state includes an initial pressure of the gaseous phase 3G contained in the tank 3 and an initial temperature of the liquid phase 3L contained in the tank 3. This information is entered for example by a user using the acquisition means 42.

[0064] Preferably, the initial state also includes an initial composition of the liquefied gas contained in tank 3. This initial composition is generally indicated by the supplier of the liquefied gas when it is loaded into tank 3. It can, for example, be entered by the user using the acquisition means 42.

[0065] The process 300 further includes a step 302 consisting of providing a predicted route 80 of the vessel 1. This has been represented on the figure 6 As an illustration, an example of a planned route 80. As shown in this figure, the planned route 80 is defined by a starting point 81, a destination point 82, and a plurality of waypoints 83 through which the vessel 1 is expected to pass. The points 81, 82, 83 are entered by the user using the acquisition means 42. The points 81, 82, 83 are defined by their geographical coordinates in a way known per se.

[0066] Each pair of consecutive waypoints 83 together defines a journey stage 84. The planned journey 80 can further be defined by a heading to be followed by the ship 1 along the journey stages 84.

[0067] The process 300 further includes a step 303 of determining at least one environmental parameter of the vessel 1 along the predicted route 80 from meteorological forecasts. Specifically, using the communication interface 130, the central unit 110 acquires meteorological forecasts provided by a weather forecasting provider and determines at least one environmental parameter from the weather forecast and the predicted route 80. The environmental parameter includes at least one of the following: current direction, current velocity, wind speed, wind direction, mean wave height, wave direction, and wave period, and preferably several or even all of these parameters.It should be noted that some of the environmental parameters can be provided directly by the weather forecast provider, while others can be estimated by the central unit 110 by applying a model to the weather forecast.

[0068] The method 300 further includes a step 304 of determining a speed curve of the vessel 1 along the predicted route 80. In a simplified embodiment, the speed curve can simply be provided as input data, for example by the user using the acquisition means 41.

[0069] The process 300 further includes a step 305 consisting of generating a plurality of tank management scenarios 50.

[0070] There figure 7 This is a graph illustrating an example of a tank management scenario 50. As shown in this figure, a given tank management scenario 50 defines the evolution of the pressure P of the gaseous phase 3G contained in tank 3 as a function of time t, during the predicted journey. In a preferred embodiment, the predicted journey 80 is divided into consecutive time intervals d1, d2, d3, d4, d5, d6, ..., and the tank management scenario 50 defines average values ​​P1, P2, P3, P4, P5, P6, ... of the pressure P over the time intervals d1, d2, d3, d4, d5, d6, .... The time intervals d1, d2, d3, d4, d5, d6, ... can each be equal to one day.

[0071] For example, tank management scenarios 50 can be randomly generated at step 305.

[0072] After step 305, process 300 proceeds to the evaluation of tank management scenarios 50 (reference A on the figures 5A et 5B ).

[0073] According to a preferred execution mode which will be described in more detail later, process 300 implements an evolutionary algorithm using as its objective function a cost function which depends at least on a total quantity of BOG generated in tank 3 along the forecast path 80. Within the framework of this evolutionary algorithm, in a manner known per se, the tank management scenarios 50 are iteratively regenerated a plurality of times, each new generation of tank management scenarios 50 being generated from a number of tank management scenarios 50 from the previous generation which best minimize the cost function.

[0074] There figure 8 This is a block diagram that illustrates, to facilitate understanding of the description that follows, models 61, 62, and 63 that are implemented to evaluate the cost function. These models 61, 62, and 63 can be stored in database 150 of the management support system 100.

[0075] A first model 61, which will be called in the following "motor model 61", is capable of estimating a power required at the output of the motor 40, more precisely a mechanical power required at the level of an output shaft 40A of the motor 40 (see figure 2 This mechanical power may be the sole power required to propel vessel 1. Alternatively, it may be the sum of the power required to propel vessel 1 and the power required to supply electrical power to other equipment on vessel 1 (commonly referred to as "Hotel Load"). In any case, the engine model 61 is capable of estimating the power required at the output of engine 40 based on a speed setpoint for vessel 1, at least one environmental parameter, and possibly the heading of vessel 1.

[0076] According to a specific implementation method, the engine model 61 is a statistical model trained using a supervised machine learning method. This statistical model is trained on a set of test data, which can be obtained from navigation tests consisting of sailing a vessel and measuring: on the one hand, a ship speed and at least one environmental parameter chosen from wave height, wave period, angle between wave direction and ship heading, wind speed, angle between wind direction and ship heading; on the other hand, a power output from the ship's engine required to maintain the ship's speed.

[0077] A second model 62, which we will subsequently call the "tank model 62", is capable of estimating the quantity of BOG generated in tank 3. The tank model 62 estimates the quantity of BOG generated in the tank by separately estimating: a first quantity of BOG generated in tank 3 as a function of the current pressure P of the gaseous phase 3G contained in tank 3 and the current temperature T of the liquid phase 3L contained in tank 3; and a second quantity of BOG generated in tank 3 as a function of the current pressure P of the gaseous phase 3G and at least one environmental parameter, then by adding these two quantities of BOG.

[0078] The first quantity of BOG reflects the generation of BOG linked to the coexistence of the liquid phase 3L and the gaseous phase 3G. It can be estimated by a thermodynamic model known in itself, for example a Hertz-Knudsen model.

[0079] The second quantity of BOG reflects the additional generation of BOG in tank 3 due to the agitation of the 3L liquid phase caused by the movements undergone by vessel 1.

[0080] According to a specific execution method, the second BOG quantity is estimated by a statistical model trained using a supervised machine learning method. This statistical model is trained on a test dataset, which can be obtained from navigation tests consisting of sailing a vessel and measuring: on the one hand, a pressure of the gaseous phase contained in the tank and a temperature of the liquid phase contained in the tank; and on the other hand, a difference between the quantity of BOG actually generated in the ship's tank and an estimated quantity of BOG estimated by a thermodynamic model.

[0081] Furthermore, if, as mentioned above, an initial composition of the liquefied gas contained in tank 3 is provided in step 301, the tank model 62 can further estimate the composition of the BOG generated in tank 3. In this case, the model estimating the first quantity of BOG and the model estimating the second quantity of BOG each estimate the composition of the BOG generated in tank 3.

[0082] A third model 63, which will be referred to hereafter as "flow model 63", is capable of estimating the quantity of gas to be extracted from tank 3 by the vapor phase treatment subsystem 13 and the quantity of BOG to be burned in the GCU 30. More specifically, the flow model 63 estimates, based on the power estimated by model 61, the quantity of BOG to be supplied to the engine 40, and compares this quantity of BOG to be supplied to the engine 40 with the quantity of BOG generated in tank 3 estimated by the tank model 62. If the quantity of BOG to be supplied to the engine 40 is less than the quantity of BOG generated in tank 3, the flow model 63 determines that the difference between these two quantities is to be burned in the GCU 30.Conversely, if the amount of BOG to be supplied to the engine 40 is greater than the amount of BOG generated in the tank 3, the flow model 63 determines that the difference between these two amounts is to be extracted from the tank 3 by the vapor phase treatment subsystem 13.

[0083] Models 61, 62, 63 are implemented iteratively, with a very small time step compared to the duration of the intervals d1, d2, ..., so as to estimate at each iteration the pressure P of the gaseous phase 3G, the temperature T of the liquid phase 3L, the quantity of BOG generated by the tank 3 (and where applicable the composition of the BOG generated by the tank 3), and the quantity of BOG to be supplied to the engine 40 at the GCU 30.

[0084] Returning to the figure 5B We will now describe the steps involved in evaluating the cost function. N being the number of tank management scenarios 50 generated in step 305, process 300 performs N evaluations S1, S2, ..., SN for each of the tank management scenarios 50. Although the diagram of the figure 5B The S1, S2, ..., SN assessments are presented as being carried out sequentially, although they may also be carried out in parallel. Each S1, S2, ..., SN assessment comprises steps 401, 402, 403, and 404, which are detailed below.

[0085] At step 401, from the speed curve of vessel 1 determined in step 304 and at least one environmental parameter determined in step 303, a power curve required at the output of engine 40 along the predicted route 80 is estimated using the engine model 61.

[0086] In step 402, a BOG quantity curve generated in tank 3 along the predicted path 80 is estimated using tank model 62.

[0087] In step 403, from the power curve required at the output of engine 40 estimated in step 401 and the curve of quantity of BOG generated in tank 3 estimated in step 402, a curve of quantity of gas to be extracted from tank 3 and a curve of quantity of BOG to be burned in GCU 30 are estimated using the flow model 63.

[0088] Following steps 401 to 403, in step 404, a cost function is calculated. More specifically, the cost function is a real number calculated from the estimates made in steps 401 to 403, and which quantifies the acceptability of the tank 50 management scenario in question.

[0089] In a simple execution mode, the cost function depends only on the total quantity of BOG generated in tank 3 along the predicted path 80. However, it is preferable that the cost function also depend on other quantities or criteria.

[0090] In a preferred execution mode, the cost function also depends on: of a difference between the pressure P of the gaseous phase contained in tank 3 at the end of the planned route 80 and the discharge pressure required by a discharge terminal located at the destination point 82; and / or of a difference between the temperature T of the liquid phase contained in tank 3 at the end of said planned route 80 and a discharge temperature required by a discharge terminal located at the destination point 82; and preferably both of these differences. In this way, a tank management scenario 50 is considered all the more acceptable as it allows arrival at the discharge terminal located at the destination point 82 with a discharge pressure and a discharge temperature close to those required by the discharge terminal.

[0091] In addition or as an alternative, the cost function may also depend on whether the tank management scenario 50 meets some or all of the following criteria: the pressure P of the gaseous phase contained in tank 3 remains below a maximum safety pressure; this maximum safety pressure may be a fixed value defined in advance, or a variable value which depends on environmental conditions encountered by the vessel 1 along the planned route 80, and which may, for example, be lower in the event of a storm or other potentially dangerous navigation conditions; the pressure P of the gaseous phase contained in tank 3 remains above a minimum pressure; this minimum pressure may be a fixed value defined in advance, or a variable value which depends on the amount of BOG required by the engine 40, as estimated by the flow model 63; the flow rate of BOG through the vapor phase treatment system 10 remains below a threshold value.

[0092] As previously mentioned, each of steps 401 to 404 is carried out within each of the S1, S2, ..., SN assessments of tank management scenarios 50.

[0093] After evaluations S1, S2, ..., SN, process 300 proceeds to step 306, which consists of verifying whether a stopping criterion is met. In one execution mode, the stopping criterion is a computation time criterion. In other words, the stopping criterion is considered met if a predefined computation time is reached or exceeded. Alternatively, the stopping criterion can be a number of iterations criterion.

[0094] In any case, if the stopping criterion is not met (reference "N" in 306 on the figure 5B In step 300, process 300 proceeds to a step 307 consisting of selecting k tank 50 management scenarios from among the N tank 50 management scenarios evaluated during evaluations S1, S2, ..., SN, where k < N. The k tank 50 management scenarios selected in step 307 are those that best minimize the cost function calculated in step 404. In one variant, k is a predetermined number. In another variant, k is not predetermined, and the k tank 50 management scenarios selected are those for which the cost function calculated in step 404 is less than a threshold.

[0095] After step 307, process 300 proceeds to a step 308 which consists of generating N new tank 50 management scenarios from the k tank 50 management scenarios selected in step 307.

[0096] As mentioned above, in this execution mode, process 300 implements an evolutionary algorithm using as its objective function the cost function calculated in step 404. In a particular execution mode, the evolutionary algorithm is a genetic algorithm. Optimization methods using an evolutionary algorithm are well known as such. Step 305 then consists of initializing the population considered by the evolutionary algorithm by randomly generating N tank management scenarios, and step 308 consists of generating N new tank management scenarios by crossing and mutating the tank management scenarios selected in step 307.

[0097] If, on the other hand, the stopping criterion is met (reference "O" in 306 on the figure 5B The process 300 proceeds to a step 309 consisting of selecting p tank management scenarios from among the N tank management scenarios evaluated during evaluations S1, S2, ..., SN, where 1 ≤ p < N. The p tank management scenarios selected in step 307 are those that best minimize the cost function calculated in step 404. In one variant, p is a predetermined number. In another variant, p is not predetermined, and the p selected tank management scenarios are those for which the cost function calculated in step 404 is less than a threshold.

[0098] After step 309, process 300 proceeds to a step 310 which consists of displaying to the user the tank management scenarios selected in step 309, for example on the display means 41.

[0099] In a highly simplified alternative execution mode, a single tank management scenario 50 is generated in step 305. Steps 401 to 404 are then performed for this single tank management scenario 50, and steps 306, 307, and 308 are omitted. In step 309, it is determined, based on the cost function calculated in step 404, whether the tank management scenario 50 is acceptable. If so, this tank management scenario 50 is displayed in step 310.

[0100] In another simplified alternative execution mode, a plurality of tank management scenarios 50 are generated at step 308, but steps 306, 307, 308 are omitted.

[0101] The steps of process 300 can be repeated at fixed intervals and / or upon receipt of new weather forecasts via communication interface 130. In addition or as an alternative, the steps of process 300 can be repeated on user command.

[0102] In any event, process 300 leads, at step 310, to displaying to the user, for example on the display means 41, at least one tank management scenario 50, according to the cost function calculated in step 404. Process 300 thus makes it possible to offer decision support to the user.

[0103] As mentioned above, a given tank management scenario 50 can define average values ​​P1, P2, P3, P4, P5, P6, ... of the pressure P over consecutive time intervals d1, d2, d3, d4, d5, d6, .... In a preferred execution mode, after step 309 and before step 310, the tank management scenario(s) 50 selected in step 1 can be post-processed so as to further display to the user minimum and maximum values ​​of the pressure P over consecutive time intervals d1, d2, d3, d4, d5, d6, ... .

[0104] There figure 9 This is a graph illustrating an example of a tank management scenario 50 that has been post-processed. As shown in this figure, the tank management scenario 50 defines, over the interval d1, in addition to the average value P1, a maximum value M1 and a minimum value m1, with the pressure P being neither greater than the maximum value M1 nor less than the minimum value m1. Similarly, over each interval d2, d3, d4, d5, d6, ..., the tank management scenario 50 defines, respectively, maximum values ​​M2, M3, M4, M5, M6, ... and minimum values ​​m2, m3, m4, m5, m6, ....

[0105] The maximum value M1 is calculated based on the average value P1 and one or more constraints relating to tank 3. Constraints relating to tank 3 may include safety constraints for tank 3, such as the maximum safe pressure of the tank. All or some of the constraints relating to tank 3 may be entered by the user using the acquisition means 42.

[0106] The maximum values ​​M2 to M6 are calculated in a similar way, based on the average values ​​P2 to P6 and one or more constraints relating to tank 3.

[0107] The minimum value m1 is calculated based on the average value P1 and the quantity of gas to be extracted from tank 3 during the interval d1 estimated in step 403 using the flow model 63.

[0108] The minimum values ​​m2 to m6 are calculated in a similar way, based on the average values ​​P2 to P6 and the quantities of gas to be extracted from tank 3 during the intervals d2 to d6.

[0109] In a simple implementation variant, the post-processed tank management scenario(s) 50 can simply be displayed to the user, i.e., shown together with the corresponding minimum values ​​m1 to m6 and maximum values ​​M1 to M6. This provides the user with additional decision support by also displaying a range of values, i.e., the interval [m1, M1], [m2, M2], ... within which the pressure P can be allowed to vary during the interval d1, d2, ....

[0110] Alternatively, either only the maximum values ​​M1 to M6, or only the minimum values ​​m1 to m6 can be calculated and then displayed.

[0111] In a particularly preferred variant, the tank management scenario(s) 50 thus post-treated may also go, before step 310, through an additional optimization step aimed at determining optimal values ​​of the pressure P over sub-intervals of the intervals d1, d2, ....

[0112] There figure 10 is a graph illustrating an example of such an optimized 50A tank management scenario. As shown in this figure, the time interval d1 is subdivided into several consecutive sub-time intervals i1, i2, i3, i4, i5, i6.... The time intervals i1, i2, i3, i4, i5, i6... can each be equal to one hour. In addition to the average value P1 and the minimum and maximum values ​​m1, the 50A tank management scenario defines average values ​​Pi1, Pi2, Pi3, Pi4, Pi5, Pi6 for each of the sub-time intervals i1, i2, i3, i4, i5, i6. Similarly, and although not shown in the figure 10 For the sake of brevity, the time intervals d2, d3, d4, d5, d6, ... are also subdivided into several consecutive sub-intervals of time, an average value of the pressure P being defined on each of these sub-intervals.

[0113] The average values ​​Pi1, Pi2, ... are determined by an evolutionary algorithm using as its objective function a cost function which depends at least on a total quantity of BOG generated in tank 3 during the interval i1 and on compliance with the following criteria: the pressure P of the gaseous phase contained in tank 3 remains below the maximum value M1; the pressure P of the gaseous phase contained in tank 3 remains above the minimum value m1; the flow rate of BOG through the vapor phase treatment system 10 remains below a threshold value.

[0114] This determination is carried out through steps similar to those in steps 401 to 404 and 306 to 308, which are therefore not described in detail again. The 50A tank management scenario(s) are then displayed to the user.

[0115] As mentioned above, the velocity curve can simply be provided as input data, for example by the user using the acquisition means 41, at step 304. However, it is preferable, at step 304, for the velocity curve to be determined by an evolutionary algorithm using an objective function that depends on a total amount of BOG generated in tank 3 along the predicted route 80 and a difference between a required time of arrival at the destination of vessel 1 and an estimated time of arrival at the destination of vessel 1. The required time of arrival at the destination of vessel 1 can be entered by the user using the acquisition means 42.

[0116] Within the framework of this evolutionary algorithm, in a manner known per se, 550 speed management scenarios are iteratively generated a plurality of times, each new generation of 550 speed management scenarios being generated from a number of 550 speed management scenarios from the previous generation that best minimize the objective function.

[0117] There figure 11 This is a graph illustrating an example of a speed management scenario 550. As shown in this figure, the planned route 80 is divided into consecutive intervals of distance traveled y1, y2, y3, y4, y5, ..., a given speed management scenario 550 defines an evolution of the speed V of vessel 1 by defining average values ​​V1, V2, V3, V4, V5, ... of the speed V over each of the consecutive intervals of distance traveled y1, y2, y3, y4, y5, .... The distance intervals traveled y1, y2, y3, y4, y5, ... can advantageously be defined by reference to points 81, 82, 83 of the planned route 80.

[0118] With reference to the figure 12 We will now describe the implementation steps of the evolutionary algorithm.

[0119] During a step 601, a plurality of speed management scenarios 550 are generated, for example randomly.

[0120] After step 601, where M is the number of speed management scenarios 550 generated in step 601, M evaluations R1, R2, ..., RM are performed for each of the speed management scenarios 550. Although the diagram of the figure 12 represents the R1, R2, ..., RM evaluations as being carried out sequentially; these may optionally be carried out in parallel. Each R1, R2, ..., RM evaluation comprises steps 602, 603, 604, 605, 606, which are detailed below.

[0121] In step 602, an effective velocity curve of vessel 1 is estimated using a hydrodynamic model of vessel 1.

[0122] In step 603, an effective trajectory of vessel 1 is estimated from the effective velocity curve estimated in step 602 and using an ocean current model.

[0123] In step 604, at least one environmental parameter of vessel 1 is determined along the actual trajectory of vessel 1 from weather forecasts. Specifically, similarly to step 303, using the communication interface 130, the central unit 110 acquires weather forecasts provided by a weather forecast provider and determines at least one environmental parameter from the weather forecast and the actual trajectory of vessel 1 estimated in step 603.

[0124] In step 605, a BOG quantity curve generated in tank 3 along the effective trajectory of vessel 1 is estimated using tank model 62.

[0125] After steps 601 to 605, in step 606, the objective function is calculated. The objective function is a real number that quantifies the acceptability of the speed management scenario 550. As mentioned above, the objective function depends on a total amount of BOG generated in tank 3 along the planned route 80 and a difference between a required time of arrival at the destination of vessel 1 and an estimated time of arrival at the destination of vessel 1.

[0126] After evaluations R1, R2, ..., RM, we proceed to step 607, which consists of verifying whether a stopping criterion is met. In one execution mode, the stopping criterion is a computation time criterion. In other words, the stopping criterion is considered met if a predefined computation time is reached or exceeded. Alternatively, the stopping criterion can be a number of iterations criterion.

[0127] In any case, if the stopping criterion is not met (reference "N" in 607 on the figure 12 ), we proceed to step 608, which consists of selecting k' speed management scenarios 550 from among the M speed management scenarios 550 evaluated during evaluations R1, R2, ..., RM, where k' < M. The k' speed management scenarios 550 selected in step 608 are those that best minimize the objective function calculated in step 606. In one variant, k' is a predetermined number. In another variant, k is not predetermined, and the k' speed management scenarios 550 selected are those for which the objective function calculated in step 608 is less than a threshold.

[0128] After step 608, we move on to step 609 which consists of generating M new tank management scenarios from the k' speed management scenarios 550 selected in step 608.

[0129] In a particular execution mode, the evolutionary algorithm is a genetic algorithm. Step 601 then consists of initializing the population considered by the evolutionary algorithm by randomly generating M speed management scenarios 550, and step 609 consists of generating M new speed management scenarios 550 by crossing and mutating the speed management scenarios 550 selected in step 608.

[0130] If, on the other hand, the stopping criterion is met (reference "O" in 607 on the figure 12 ), we proceed to step 610, which consists of selecting the speed management scenario 550 that best minimizes the objective function calculated in step 606. Finally, in step 611, this speed management scenario 550 is designated as the ship speed curve 1 along said forecast route, after which steps 305 to 310 and the evaluations S1, S2, ..., SN can be implemented as already described above.

[0131] The configuration of the vapor phase treatment system 10 shown on the figure 2 This is just one example. Other configurations are possible.

[0132] In some variants not shown, the vessel 1 may have more than one engine 40. In this case, the engine model 61 is able to estimate the power required at the output of each engine 40 as a function of a speed setpoint of the vessel 1, of at least one environmental parameter, and possibly of the heading of the vessel 1.

[0133] As mentioned above in relation to the figure 2 The engine 40 can provide both the power required to propel Ship 1 and the power required to supply electrical power to other equipment on Ship 1 (commonly referred to as "Hotel Load"). However, as shown on the figure 13 The vessel 1 may include an auxiliary engine 49 intended to provide this "Hotel Load" power. The auxiliary engine 49, like the engine 40, is capable of consuming BOG from the BOG distribution unit 20 or BOG from the heater 60. In this case, the engine model 61 is capable of estimating the power required at the output of the engine 40 and the auxiliary engine 49 as a function of a speed setpoint for the vessel 1, at least one environmental parameter, and possibly the heading of the vessel 1. It is further specified that, alternatively, several auxiliary engines 49 may be provided; in this case, similarly, the engine model 61 is capable of estimating the power required at the output of the engine(s) 40 and the auxiliary engines 49.

[0134] We have represented on the figure 14 another variant of the vapor phase treatment system 10, which also includes a reliquefaction plant 75.

[0135] As is known per se, the reliquefier 75 is capable of receiving BOG from the BOG distribution unit 20, reliquefying it, and sending it back to tanks 3, 4, 5, 6. Such a reliquefier 75 is known as such and is therefore not described in detail here.

[0136] It is specified that in a way not represented on the figure 14 The power (for example, electrical power) required for the operation of the reliquefactor 75 can be supplied by the motor 40 or by the auxiliary motor 49. It is further specified that in certain variants not shown, the reliquefactor 75 may send reliquefied BOG only to one or some of the tanks 3, 4, 5, 6.

[0137] Of course, in this variant, the flow model 63 takes into account the presence of the reliquefactor 75.

[0138] More specifically, the flow model 63 is capable of estimating the quantity of gas to be extracted from tank 3 by the vapor phase treatment subsystem 13, the quantity of BOG to be burned in the GCU 30, and furthermore, the quantity of BOG to be sent to the reliquefaction unit 75. More specifically, the flow model 63 estimates, based on the power estimated by model 61, the quantity of BOG to be supplied to the engine 40, and compares this quantity of BOG to be supplied to the engine 40 with the quantity of BOG generated in tank 3 estimated by the tank model 62. If the quantity of BOG to be supplied to the engine 40 is greater than the quantity of BOG generated in tank 3, the flow model 63 determines that the difference between these two quantities must be extracted from tank 3 by the vapor phase treatment subsystem 13.Conversely, if the quantity of BOG to be supplied to engine 40 is less than the quantity of BOG generated in tank 3, the flow model 63 determines that the difference between these two quantities is to be sent preferentially to the reliquefactor 75, and furthermore to be burned in the GCU 30 only if the reliquefaction capacity of the reliquefactor 75 is insufficient.

[0139] It is further understood that in this variant, at step 403, from the power curve required at the output of engine 40 estimated in step 401 and the curve of quantity of BOG generated in tank 3 estimated in step 402, a curve of quantity of gas to be extracted from tank 3, a curve of quantity of BOG to be burned in GCU 30, and a curve of quantity of BOG to be sent to reliquefaction unit 75 are estimated using the flow model 63.

[0140] We have represented on the figure 15 yet another variant of the vapor phase treatment system 10, which also includes a subcooler 77 (“subcooler” in English).

[0141] As is known per se, the subcooler 77 is suitable for receiving from the vapor phase treatment subsystem 13 a portion of the liquid phase 3L contained in the tank 3 and for lowering its temperature (for example, by about 5 to 10 degrees Celsius, i.e., from about -162°C to about -167°C to -172°C in the case where the liquefied gas is LNG at atmospheric pressure). The subcooler 77 is further suitable for returning the subcooled liquid phase thus subcooled to the tank 3, either directly into the liquid phase 3L by means of a mixing nozzle 79, known as such, or indirectly by spraying into the gaseous phase 3G in the form of droplets by means of a spraying ramp 78.

[0142] It is specified that in a way not represented on the figure 15 The power (for example, electrical power) required for the operation of the subcooler 77 can be supplied by the motor 40 or by the auxiliary motor 49. It is further specified that in certain variants not shown, the subcooler 77 may be able to extract the liquid phase from several or even all of the tanks 3, 4, 5, 6 and return it subcooled to the respective tank.

[0143] Of course, in this variant, the flow model 63 takes into account the presence of the subcooler 77.

[0144] More specifically, the flow model 63 is capable of estimating the quantity of gas to be extracted from tank 3 by the vapor phase treatment subsystem 13, the quantity of BOG to be burned in the GCU 30, and also the quantity of liquid phase to be sent to the subcooler 77. More specifically, the flow model 63 estimates, based on the power estimated by model 61, the quantity of BOG to be supplied to the engine 40, and compares this quantity of BOG to be supplied to the engine 40 with the quantity of BOG generated in tank 3 estimated by the tank model 62. If the quantity of BOG to be supplied to the engine 40 is greater than the quantity of BOG generated in tank 3, the flow model 63 determines that the difference between these two quantities must be extracted from tank 3 by the vapor phase treatment subsystem 13.Conversely, if the amount of BOG to be supplied to the engine 40 is less than the amount of BOG generated in the tank 3, the flow model 63 determines that the difference between these two amounts is to be reduced over time by preferentially operating the subcooler 77, the possible excess of BOG being burned in the GCU 30 only if the subcooling capacity of the subcooler 77 is insufficient.

[0145] It is further understood that in this variant, at step 403, from the power curve required at the output of engine 40 estimated in step 401 and the curve of quantity of BOG generated in tank 3 estimated in step 402, a curve of quantity of gas to be extracted from tank 3, a curve of quantity of BOG to be burned in the GCU 30, and a curve of quantity of liquid phase to be extracted from tank 3 and sent to subcooler 77 are estimated using the flow model 63.

[0146] Finally, it is specified that in certain variants not shown, the subcooler 77 may be able to extract the liquid phase from one of the tanks 3, 4, 5, 6 and return it subcooled to one or more other tanks 3, 4, 5, 6. In this case, when the management aid process 300 is implemented simultaneously for several or each of the tanks 3, 4, 5, 6 as mentioned above, the flow model 63 relating to a given tank further estimates a quantity of subcooled liquid phase to be received from the subcooler 77 from other tank(s).

[0147] Although the invention has been described in connection with several particular embodiments, it is clearly evident that it is by no means limited to them and that it includes all technical equivalents of the means described as well as their combinations if these fall within the scope of the invention.

[0148] The use of the verb "comporter", "comprendre" or "include" and its conjugated forms does not exclude the presence of other elements or steps than those stated in a claim.

[0149] In claims, any reference sign in parentheses should not be interpreted as a limitation of the claim.

Claims

1. A computer-implemented method for assisting the management of a vessel (1) for transporting liquefied gas, the vessel comprising at least one tank (3) configured to contain liquefied gas and a vapor phase treatment system (10), the vapor phase treatment system (10) being capable of sending boil-off gas exiting the tank (3) to a propulsion engine (40) of the vessel or to a gas combustion unit (30) on board the vessel, the vapor phase treatment system (10) further being capable of extracting a portion of the liquid phase (3L) contained in the tank (3) and of evaporating this portion in order to send it to the propulsion engine (40), the method comprising: - providing (301) an initial state of the tank, said initial state of the tank comprising an initial pressure of the gas phase (3G) contained in the tank (3) and an initial temperature of the liquid phase (3L) contained in the tank (3); - providing (302) a forecast path (80) of the vessel (1); - determining (303), on the basis of weather forecasts, at least one environmental parameter of the vessel along said forecast path (80); - determining (304) a speed curve of the vessel (1) along said forecast path (80); the method being characterized in that it comprises: - generating (305) a tank management scenario (50; 50A) defining an evolution of the pressure of the gas phase (3G) contained in the tank during said forecast path (80), and, on the basis of the tank management scenario thus generated: a) estimating (401) a curve of the power required to be output from the propulsion engine (40) along said forecast path (80) as a function of said speed curve of the vessel and of said at least one environmental parameter; b) estimating (402) a curve of the amount of boil-off gas generated in the vessel (3) along said forecast path; c) estimating (403), on the basis of said required power curve and of said curve of the amount of boil-off gas generated, a curve of the amount of gas to be extracted from the tank and a curve of the amount of boil-off gas to be burned in the gas combustion unit (30); and d) computing (404) a cost function that at least depends on a total amount of boil-off gas generated in the tank along said forecast path (80); and - displaying (310) to a user the tank management scenario (50; 50A) as a function of the computed cost function.

2. The method as claimed in claim 1, wherein multiple tank management scenarios (50; 50A) are generated, with steps a) to d) being carried out for each tank management scenario, and wherein at least one of said tank management scenarios is displayed to the user as a function of the computed cost function for said tank management scenario.

3. The method as claimed in claim 2, wherein, before displaying (310) at least one of said tank management scenarios (50; 50A) to the user, said tank management scenarios are iteratively regenerated a number of times by a first evolutionary algorithm using the cost function as a first objective function.

4. The method as claimed in any one of claims 1 to 3, wherein said forecast path (80) comprises path steps (84) each defined by two waypoints (83) and a heading to be followed between said two waypoints, and wherein, in step a), the curve of the power required to be output from the propulsion engine (40) along said forecast path (80) is estimated as a function of said headings to be followed, of said speed curve of the vessel and of said at least one environmental parameter.

5. The method as claimed in any one of claims 1 to 4, wherein said speed curve of the vessel is determined from said forecast path (80) and from said at least one environmental parameter by a second evolutionary algorithm using a second objective function that depends on said total amount of boil-off gas generated in the tank along said forecast path (80) and on a difference between a required time of arrival at the destination and an estimated time of arrival at the destination.

6. The method as claimed in any one of claims 1 to 5, wherein providing (302) a forecast path (80) of the vessel comprises providing an offloading pressure and / or an offloading temperature required by an offloading terminal at the destination of the vessel, and wherein, in step d), the cost function further depends on a difference between the pressure of the gas phase (3G) contained in the tank (3) at the end of said forecast path (80) and said offloading pressure and / or a difference between the temperature of the liquid phase (3L) contained in the tank (3) at the end of said forecast path (80) and said offloading temperature.

7. The method as claimed in any one of claims 1 to 6, wherein, in step a), the required power curve is estimated using a first statistical model, the first statistical model being capable of estimating a power required to be output from the propulsion engine (40) at least as a function of a vessel speed setpoint and of said at least one environmental parameter, and the first statistical model being trained by a supervised machine learning method.

8. The method as claimed in any one of claims 1 to 7, wherein the aforementioned step b) involving estimating (402) a curve of the amount of boil-off gas generated in the tank (3) along said forecast path (80) comprises estimating a first amount of boil-off gas generated in the tank as a function of a current pressure of the gas phase (3G) contained in the tank and of a current temperature of the liquid phase (3L) contained in the vessel and estimating a second amount of boil-off gas generated in the tank as a function of a current pressure of the gas phase (3G) contained in the tank and of said at least one environmental parameter.

9. The method as claimed in claim 8, wherein the second amount of boil-off gas generated in the tank (3) is estimated using a second statistical model, the second statistical model being trained by a supervised machine learning method.

10. The method as claimed in any one of claims 1 to 9, wherein the initial state comprises an initial composition of the liquefied gas contained in the tank, and wherein step b) involving estimating (402) a curve of the amount of boil-off gas generated in the tank (3) along said forecast path (80) further comprises estimating a composition of the boil-off gas generated in the tank.

11. The method as claimed in any one of claims 1 to 10, wherein the environmental parameter comprises at least one from among a current direction, a current speed, a wind speed, a wind direction, an average wave height, a wave direction, and a wave period.

12. The method as claimed in any one of claims 1 to 11, wherein the vapor phase treatment system (10) further comprises a reliquefaction plant (75) capable of reliquefying boil-off gas exiting the tank (3) and of returning the boil-off gas thus re-liquefied to the tank, and wherein the aforementioned step c) comprises estimating (403), on the basis of said required power curve and of said curve of the amount of boil-off gas generated, a curve of the amount of gas to be extracted from the tank, a curve of the amount of boil-off gas to be burned in the gas combustion unit (30), and a curve of the amount of boil-off gas to be sent to the reliquefaction plant (75).

13. The method as claimed in any one of claims 1 to 12, wherein the vapor phase treatment system (10) further comprises a sub-cooler (77) capable of sub-cooling a portion of the liquid phase (3L) contained in the vessel (3) and of returning the portion thus sub-cooled to the tank, and wherein the aforementioned step c) comprises estimating (403), on the basis of said required power curve and of said curve of the amount of boil-off gas generated, a curve of the amount of gas to be extracted from the tank, a curve of the amount of boil-off gas to be burned in the gas combustion unit (30), and a curve of the amount of liquid phase to be extracted from the tank and to be sent to the sub-cooler (77).

14. The method as claimed in any one of claims 1 to 13, wherein the steps of the method as claimed in claim 1 are repeated at fixed intervals and / or upon receipt of new weather forecasts.

15. A system (100, 200) for assisting the management of a vessel for transporting liquefied gas, the vessel (1) comprising at least one tank (3) configured to contain liquefied gas and a vapor phase treatment system (10), the vapor phase treatment system (10) being capable of sending boil-off gas exiting the tank (3) to a propulsion engine (40) of the vessel or to a gas combustion unit (30) on board the vessel, the vapor phase treatment system (10) further being capable of extracting a portion of the liquid phase (3L) contained in the tank (3) and of evaporating this portion in order to send it to the propulsion engine (40), the management assistance system (100, 200) comprising: - an acquisition unit (42) configured to acquire an initial state of the tank and a forecast path (80) of the vessel, said initial state of the tank comprising an initial pressure of the gas phase (3G) contained in the tank (3) and an initial temperature of the liquid phase (3L) contained in the tank (3); - a first communication unit (130) configured to obtain weather forecasts from a weather forecast provider; the system for assisting the management being characterized in that it comprises: - a computation unit (110) configured to: on the basis of weather forecasts obtained by the first communication unit (130), determine at least one environmental parameter of the vessel along said forecast path (80); determine (304) a speed curve of the vessel along said forecast path (80); generate (305) a tank management scenario (50; 50A) defining an evolution in the pressure of the gas phase (3G) contained in the tank (3) along said forecast path (80), and, on the basis of the tank management scenario thus generated, configured to: a) estimate (401) a curve of the power required to be output from the propulsion engine (40) along said forecast path (80) as a function of said speed curve of the vessel and of said at least one environmental parameter; b) estimate (402) a curve of the amount of boil-off gas generated in the tank (3) along said forecast path; c) estimate (403), on the basis of said required power curve and of said curve of the amount of boil-off gas generated, a curve of the amount of gas to be extracted from the vessel and a curve of the amount of boil-off gas to be burned in the gas combustion unit (30); d) compute (404) a cost function that at least depends on a total amount of boil-off gas generated in the tank along said forecast path (80); and comprising - a display unit (41) configured to display to a user the tank management scenario (50; 50A) as a function of the computed cost function.

16. The system (100, 200) as claimed in claim 15, wherein said forecast path (80) comprises path steps (84) each defined by two waypoints (83) and a heading to be followed between said two waypoints, and wherein the computation unit (110) is configured to: a) estimate (401) a curve of the power required to be output from the propulsion engine (40) along said forecast path (80) as a function of said headings to be followed, of said speed curve of the vessel and of said at least one environmental parameter.

17. The system (200) as claimed in claim 15 or 16, wherein the acquisition unit (42) and the display unit (41) are on board the vessel, the first communication unit (130) and the computation unit (110) are located in a ground station (1000), and wherein the system further comprises a second communication unit (230) configured to connect the acquisition unit (42) and the computation unit (110).