A method and system for predicting ship fuel consumption

Abstract

The present invention provides a method and system for predicting ship fuel consumption. First, ship model test data measured by ship model tests is obtained, and ship data, meteorological data, and navigation state data during ship navigation are respectively collected at regular intervals. All the collected data is divided into multiple data sets at a certain time interval. Then, each data set is screened to select the data sets that meet the preset conditions. A specific calculation method is used to calculate the wind resistance increase, the real ship still water resistance, and the propeller open water efficiency, and a BP neural network is used to predict the wave resistance increase. Furthermore, the main engine shaft power is calculated. Then, the value of the main engine fuel consumption rate curve function is obtained based on the main engine shaft power. Finally, the predicted main engine fuel consumption is calculated according to the main engine shaft power and the value of the main engine fuel consumption rate curve function, effectively improving the accuracy of ship fuel consumption prediction, so as to facilitate the crew to optimize the ship speed and route.

Description

Technical Field

[0001] The present invention relates to the technical field of ship fuel consumption estimation, and particularly relates to a ship fuel consumption estimation method and system. Background Art

[0002] With the increasingly prominent environmental problems, issues such as global warming caused by the greenhouse gas effect, glacier melting in polar regions, and ecological environment damage are urging people to maintain the prosperity of the economic society in a sustainable development manner. With the rising fuel prices, increasing port fees, etc., the operating costs of ships have increased significantly, and the shipping industry is facing a fierce competitive environment. Under the condition of meeting regulatory requirements, strengthening the control of ship operating costs and optimizing the ship speed and route have become the key to whether shipping enterprises can win in the market competition. However, the optimization of ship speed and route will involve the fuel consumption estimation of operating ships, and the model structure, parameters, and estimation accuracy of fuel consumption estimation are one of the difficult problems that are difficult to overall balance in the industry at present.

[0003] At present, there are mainly three means and technologies for ship fuel consumption estimation at home and abroad: The first is to combine the experimental data of the target ship model, the resistance calculation formula, and the empirical formula for fuel consumption estimation under meteorological conditions. However, this method requires a large amount of basic data and the basic data needs to be relatively accurate, especially the performance data of the ship model. Generally speaking, the fuel consumption estimation system established by this method has low accuracy, and the mean absolute error and determination coefficient of the estimation model are not very ideal. The inaccurate accuracy of the existing empirical calculation formula for wave added resistance is one of the main reasons for the large deviation of the prediction results of this type of model; The second is to use the machine learning method to construct a black box model. The black box model has high accuracy, but the structure and parameters of this method are unknown, there is no traditional physical model of ship, engine, and propeller, and a large amount of operation data and meteorological data are required as the basis, and the quality requirements for the data set are high, and the engineering generalization ability is poor. It is necessary to collect a large amount of operation data for each target ship and clean the data for training; The third is to combine the traditional physical model calculation method of ship, engine, and propeller, and the machine learning method to construct a fuel consumption estimation grey box model. Some physical characteristic data of this model can be directly measured, and the other characteristics that are difficult to obtain are obtained through data training. Although the grey box model can couple the advantages of white box and black box calculations, the difficulty of constructing such a coupled model is relatively large, and there are also few achievements and products in this regard on the market.

[0004] Therefore, there is an urgent need for a ship fuel consumption estimation method with high accuracy, good safety, economy, and reliability. Summary of the Invention

[0005] To solve the problems of low accuracy, economy, and reliability in the current evaluation process of ship fuel consumption models, the present invention provides a method for predicting ship fuel consumption. Based on ship model test data, ship data, meteorological data, and navigation state data, and using a BP neural network and a specific calculation method to calculate the predicted fuel consumption of the main engine. Based on the fusion calculation of multi-feature parameter algorithms, the accuracy of ship fuel consumption prediction is effectively improved, facilitating the crew to optimize the ship speed and route. The present invention also relates to a ship fuel consumption prediction system.

[0006] The technical solution of the present invention is as follows:

[0007] A method for predicting ship fuel consumption, characterized by comprising the following steps:

[0008] Data collection and screening step: Obtain the ship model test data measured by the ship model test, and respectively collect the ship data, meteorological data, and navigation state data during ship navigation at regular intervals. Divide all the collected data into multiple data sets at a certain time interval, and then screen each data set to select the data sets that meet the preset conditions;

[0009] Multi-feature parameter algorithm fusion calculation step, including mean calculation step, conversion step of speed through water, calculation step of wind resistance increase, calculation step of still water resistance, and prediction step of wave resistance increase,

[0010] The mean calculation step: Calculate the average value of each parameter in the ship data, meteorological data, and navigation state data respectively in the selected data sets, and sort the data sets containing the average value of each parameter in chronological order. Sequentially form new data sets by combining several consecutive data sets in the sorted data sets;

[0011] The conversion step of speed through water: In the new data set, calculate the speed through water of the ship according to the ship's speed over the ground in the navigation state data and the seawater flow rate in the meteorological data;

[0012] The calculation step of wind resistance increase: In the new data set, calculate the actual windward area of the ship according to the designed draft, designed windward area in the ship data, and the bow draft and stern draft in the navigation state data. Calculate the wind force coefficient according to the wind direction angle in the meteorological data and the bow draft and stern draft in the navigation state data. Calculate the wind resistance increase according to the wind force coefficient, the wind speed in the meteorological data, and the calculated actual windward area;

[0013] The calculation step of still water resistance: In the new data set, calculate the still water resistance of the actual ship according to the bow draft, stern draft, and ship's speed over the ground in the navigation state data;

[0014] The steps for predicting wave - added resistance: In the new dataset, calculate the propeller thrust based on the wind - added resistance and the still - water resistance of the actual ship, calculate the propeller advance speed based on the ship's speed through water, calculate the constant point based on the propeller thrust, the propeller advance speed, and the propeller diameter in the ship data, obtain the propeller advance - coefficient according to the calculation function fitted from the constant point, then calculate the open - water efficiency of the propeller under specific conditions according to the propeller advance - coefficient, calculate the actual - ship propulsion efficiency according to the open - water efficiency of the propeller, the hull efficiency, the shaft - transmission efficiency, and the relative - rotation efficiency of the propeller in the ship - model test data, and calculate the historical wave - added resistance according to the still - water resistance of the actual ship, the ship's speed through water, the wind - added resistance, the actual - ship propulsion efficiency, and the actual - ship shaft power obtained based on the ship - model shaft power in the ship - model test data. Respectively, take the relevant characteristic parameters in the meteorological data and the navigation - state data and the historical wave - added resistance as training - set samples, train the training - set samples based on the BP neural network to obtain the wave - added - resistance prediction model, and respectively take the relevant characteristic parameters in the meteorological data and the navigation - state data as inputs, and predict the wave - added resistance in a future time period according to the wave - added - resistance prediction model;

[0015] The steps for calculating the main - engine shaft power: In the new dataset, calculate the main - engine shaft power according to the wind - added resistance, the predicted wave - added resistance, the calculated still - water resistance of the actual ship, the ship's speed through water, and the actual - ship propulsion efficiency;

[0016] The steps for calculating the estimated main - engine fuel consumption: Obtain the value of the main - engine fuel - consumption rate curve function corresponding to the main - engine shaft power, and calculate the estimated main - engine fuel consumption according to the main - engine shaft power and the value of the main - engine fuel - consumption rate curve function.

[0017] Preferably, in the data - collection and screening steps, the screening of each dataset includes steady - navigation screening, shallow - water - effect screening, and steady - wind - and - wave screening;

[0018] And / or, the ship data includes the propeller diameter, the designed draft, the designed - draft windward area, and the wetted surface area of the ship, the meteorological data includes the wind speed, the wind direction, the flow direction, the wind - direction angle, the seawater flow speed, the wave - average period, the significant wave height, and the main - wave - direction angle; the navigation - state data includes the ship's bow draft, the ship's stern draft, the rudder angle, the course angle, the ship's frictional - resistance coefficient, and the speed over the ground.

[0019] Preferably, in the wave - added - resistance prediction steps, the relevant characteristic parameters include the bow draft, the stern draft, the speed over the ground, the course angle, the wave - average period, the significant wave height, and the main - wave - direction angle.

[0020] Preferably, the ship - model test includes the ship - model self - propulsion test, the ship - model wind - tunnel test, and the ship - model propeller open - water test.

[0021] Preferably, in the wave drag prediction step, calculating the open water efficiency of the propeller under specific conditions according to the propeller advance coefficient includes:

[0022] In the propeller open water characteristic curve graph constructed based on the propeller advance coefficient, constant points, and propeller open water efficiency, when the propeller advance coefficient is located at the propeller open water test data points, the propeller open water efficiency is directly obtained. When the propeller advance coefficient is located between any two adjacent propeller open water test data points, interpolation processing is performed on the propeller open water test data to obtain the propeller open water efficiency.

[0023] A ship fuel consumption estimation system, characterized in that it includes a data acquisition and screening module, a multi-feature parameter algorithm fusion calculation module, a main engine shaft power calculation module, and a main engine estimated fuel consumption calculation module connected in sequence. The multi-feature parameter algorithm fusion calculation module includes a mean calculation module, a speed through water conversion module, and a wind resistance calculation module, a still water resistance calculation module, and a wave drag prediction module that are all connected to the speed through water conversion module. The data acquisition and screening module is connected to the mean calculation module, and the wind resistance calculation module, the still water resistance calculation module, and the wave drag prediction module are all connected to the main engine shaft power calculation module;

[0024] The data acquisition and screening module obtains the ship model test data measured in the ship model test, and respectively collects the ship data, meteorological data, and navigation state data during ship navigation at regular intervals, and divides all the collected data into multiple data sets according to a certain time interval, and then screens each data set to screen out the data sets that meet the preset conditions;

[0025] The mean calculation module calculates the average value of each parameter in the ship data, meteorological data, and navigation state data respectively in the screened data sets, and sorts the data sets containing the average value of each parameter in chronological order, and successively forms new data sets by combining several consecutive data sets in the sorted data sets;

[0026] The speed through water conversion module calculates the ship's speed through water in the new data set according to the ship's speed over the ground in the navigation state data and the seawater flow rate in the meteorological data;

[0027] The wind resistance calculation module calculates the actual wetted area of the ship according to the designed draft, the wetted area facing the wind of the designed draft, the bow draft, and the stern draft in the ship data and the navigation state data in the new data set, calculates the wind force coefficient according to the wind direction angle in the meteorological data and the bow draft and the stern draft in the navigation state data, and calculates the wind resistance according to the wind force coefficient, the wind speed in the meteorological data, and the calculated actual wetted area;

[0028] Hydrostatic resistance calculation module: In the new dataset, calculate the actual ship's hydrostatic resistance based on the bow draft, stern draft, and ship's speed over the ground in the navigation state data;

[0029] The wave added resistance prediction module, in the new dataset, calculates the propeller thrust based on the wind added resistance and the actual ship's hydrostatic resistance, calculates the propeller advance speed based on the ship's speed through water, calculates the constant point based on the propeller thrust, propeller advance speed, and propeller diameter in the ship data, obtains the propeller advance coefficient according to the calculation function fitted from the constant point, then calculates the propeller open water efficiency under specific conditions based on the propeller advance coefficient, calculates the actual ship's propulsion efficiency based on the propeller open water efficiency, hull efficiency, shaft transmission efficiency, and propeller relative rotation efficiency in the ship model test data, and calculates the historical wave added resistance based on the actual ship's hydrostatic resistance, ship's speed through water, wind added resistance, actual ship's propulsion efficiency, and actual ship's shaft power obtained from the ship model shaft power in the ship model test data. Respectively, take the relevant characteristic parameters in the meteorological data and navigation state data and the historical wave added resistance as training set samples, train the training set samples based on the BP neural network to obtain the wave added resistance prediction model, and respectively take the relevant characteristic parameters in the meteorological data and navigation state data as inputs, and predict the wave added resistance in a future time period according to the wave added resistance prediction model;

[0030] The main engine shaft power calculation module, in the new dataset, calculates the main engine shaft power based on the wind added resistance, predicted wave added resistance, calculated actual ship's hydrostatic resistance, ship's speed through water, and actual ship's propulsion efficiency;

[0031] Main engine estimated fuel consumption calculation module, obtains the value of the main engine fuel consumption rate curve function corresponding to the main engine shaft power, and calculates the main engine estimated fuel consumption based on the main engine shaft power and the value of the main engine fuel consumption rate curve function.

[0032] Preferably, in the data acquisition and screening module, the screening of each dataset includes steady navigation screening, shallow water effect screening, and steady wind and wave screening;

[0033] And / or, the ship data includes propeller diameter, designed draft, designed draft windward area, and ship wet surface area, and the meteorological data includes wind speed, wind direction, flow direction, seawater flow velocity, wave average period, significant wave height, and main wave direction angle; the navigation state data includes ship bow draft, ship stern draft, rudder angle, course angle, ship frictional resistance coefficient, and speed over the ground.

[0034] Preferably, the relevant characteristic parameters include bow draft, stern draft, speed over the ground, course angle, wave average period, significant wave height, and main wave direction angle.

[0035] Preferably, the ship model test includes ship model self-propulsion test, ship model wind tunnel test, and ship model propeller open water test.

[0036] Preferably, in the wave drag prediction module, calculating the open water efficiency of the propeller under specific conditions according to the propeller advance coefficient includes:

[0037] In the propeller open water characteristic curve diagram constructed based on the propeller advance coefficient, constant points, and propeller open water efficiency, when the propeller advance coefficient is located at the propeller open water experimental data point, the propeller open water efficiency is directly obtained. When the propeller advance coefficient is located between any two adjacent propeller open water experimental data points, interpolation processing is performed on the propeller open water experimental data to obtain the propeller open water efficiency.

[0038] The beneficial effects of the present invention are:

[0039] A method for predicting ship fuel consumption provided by the present invention first obtains ship model test data measured in a ship model test, and respectively collects ship data, meteorological data, and navigation state data during ship navigation at regular intervals. All the collected data are divided into multiple data sets at a certain time interval, and then each data set is screened to select the data sets that meet the preset conditions. In the selected data sets, the average values of each parameter in the ship data, meteorological data, and navigation state data are calculated respectively, and the data sets containing the average value of each parameter are sorted in chronological order. Several consecutive data sets in the sorted data sets are sequentially combined into a new data set, and the ship speed through water and the actual wetted area facing the wind are calculated, and then the wind resistance increase is calculated, and then the still water resistance of the actual ship, the open water efficiency of the propeller, and the propulsion efficiency of the actual ship are calculated. Then, a wave resistance prediction model is trained based on the BP neural network, and the relevant characteristic parameters in the meteorological data and navigation state data are respectively used as inputs, and the wave resistance in a future time period is predicted according to the wave resistance prediction model, which can effectively improve the nonlinearity, prediction accuracy, and credibility of the model, and can be applied to other target ships of the same ship type, greatly improving the generalization degree of the model and the generalization ability of the project; then the main engine shaft power is calculated, and the value of the main engine fuel consumption rate curve function corresponding to it is obtained according to the main engine shaft power, and the predicted fuel consumption of the main engine is calculated according to the main engine shaft power and the value of the main engine fuel consumption rate curve function. The average value calculation step, the conversion step of the ship speed through water, the wind resistance increase calculation step, the still water resistance calculation step, and the wave resistance prediction step are overall equivalent to the fusion calculation of multi-characteristic parameter algorithms. That is, the present invention provides a method for predicting ship fuel consumption by fusing multi-characteristic parameter algorithms, processes the collected and screened data, and then performs the conversion of the ship speed through water, converts the influence of the flow velocity into the ship speed through water, then calculates the wind resistance increase and still water resistance during ship navigation, and uses the BP neural network to predict the non-linear wave resistance, constructs a wave resistance prediction model, and performs propulsion calculation on the resistance calculation results to calculate the main engine shaft power, and predicts the main engine fuel consumption data of the ship sailing under this meteorological condition based on the main engine shaft power, overcoming the defect of low prediction accuracy of the model caused by directly using the traditional calculation method and not considering the wave and meteorological conditions or using empirical formulas. The present invention has obvious advantages in physical structure and parameters, effectively improves the accuracy of ship fuel consumption prediction, and thus optimizes the ship speed and route.

[0040] The present invention also relates to a ship fuel consumption prediction system, which corresponds to the above-mentioned ship fuel consumption prediction method and can be understood as a system for implementing the above-mentioned ship fuel consumption prediction method. It includes a data acquisition and screening module, a multi-feature parameter algorithm fusion calculation module (mean calculation module, speed through water conversion module, wind resistance calculation module, still water resistance calculation module, wave resistance prediction module), a main engine shaft power calculation module, and a main engine predicted fuel consumption calculation module, which are connected in sequence. Each module works in coordination with each other. The data acquisition and screening module is the input module, and the mean calculation module, speed through water conversion module, wind resistance calculation module, still water resistance calculation module, and wave resistance prediction module as a whole are equivalent to the intermediate module of the multi-feature parameter algorithm fusion calculation - the multi-feature parameter algorithm fusion calculation module. The main engine shaft power calculation module and the main engine predicted fuel consumption calculation module are the output modules. By importing the ship parameters, ship navigation status information, and corresponding meteorological data at the input end into the intermediate module of the multi-feature parameter algorithm fusion calculation, the navigation data collected in real time per second can be processed, and the speed through water conversion is performed on the mean data in segments of, for example, 30 minutes arranged in chronological order, converting the influence of the flow velocity into the speed through water of the ship. Then, the wind resistance and still water resistance of the sailing ship are calculated using formulas, and the non-linear wave resistance is predicted using a BP neural network. The resistance calculation results are used for propulsion calculation to calculate the corrected still water navigation resistance and the main engine power of the ship, and the results of the corrected still water main engine power are imported into the output module. Finally, the output module predicts the main engine fuel consumption data of the sailing ship under this meteorological condition based on the main engine shaft power. The present invention can be understood as a ship fuel consumption prediction system that fuses multi-feature parameter algorithms. This system combines the advantages of existing white-box calculation and black-box calculation, and has the advantages of high accuracy, obvious physical structure and parameters. The ship fuel consumption prediction system that fuses multi-feature parameter algorithms overcomes the defects of low model prediction accuracy caused by directly using traditional calculation methods in the white-box mode, not considering wave meteorological conditions, or using empirical formulas. At the same time, it integrates machine learning algorithms and uses a BP neural network to predict wave resistance, improving the non-linearity, prediction accuracy, and credibility of the model. In addition, for the training of wave resistance by the wave resistance prediction module, its prediction function can still be applied to other target ships of the same ship type, which greatly improves the generalization degree of the model and the generalization ability of the project. Therefore, the present invention calculates the predicted fuel consumption of the main engine based on ship model test data, ship data, meteorological data, and navigation status data, and uses a BP neural network and specific calculation methods, effectively increasing the accuracy and credibility of ship fuel consumption prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is a flowchart of the ship fuel consumption prediction method of the present invention.

[0042] Figure 2 is a structural block diagram of the data screening of the present invention.

[0043] Figure 3 is the open water characteristic curve of the propeller of the present invention.

[0044] Figure 4 is the structural block diagram of the wave added resistance prediction of the present invention.

[0045] Figure 5 is the structural schematic diagram of the BP neural network of the present invention. Detailed implementation manners

[0046] The present invention will be described below with reference to the accompanying drawings.

[0047] The present invention relates to a method for predicting ship fuel consumption. The flowchart of the method is as Figure 1 shown, and successively includes the following steps:

[0048] 1. Data acquisition and screening step: Obtain the ship model test data measured by the ship model test, and respectively collect the ship data, meteorological data and navigation state data during ship navigation at regular intervals, and divide all the collected data into multiple data sets according to a certain time interval, and then screen each data set to screen out the data sets that meet the preset conditions. Preferably, the ship data includes the propeller diameter, designed draft, designed draft windward area and ship wet surface area; the meteorological data includes wind speed, wind direction, flow direction, wind direction angle, seawater flow velocity, wave average period, significant wave height and main wave direction angle; the navigation state data includes ship bow draft, ship stern draft, rudder angle, course angle, ship frictional resistance coefficient and speed over the ground.

[0049] Specifically, as Figure 2 shown, first obtain the ship model test data measured by the ship model self-propulsion test, wind tunnel test and propeller open water test, and respectively collect the ship data, meteorological data and navigation state data during ship navigation every 1 second, and divide all the data collected every 1 second during ship navigation into multiple 10-minute data sets according to a 10-minute time interval, and then perform calculation and screening on each 10-minute data set, such as Figure 2 shown in the stable navigation screening, shallow water effect screening and stable wind and wave screening in the data processing, and screen out the 10-minute data sets that meet the preset conditions, including:

[0050] 1.1. Stable navigation screening. For the data set of each 10-minute time period, perform the following calculation and screening:

[0051] abs(rudder angle max - rudder angle min) < 5 deg

[0052] (rotation speed max - rotation speed min) < 3 rpm

[0053] Rotation speed mean square error < 1 rpm

[0054] The mean square error of the water speed < 0.5 kn

[0055] The average flow velocity measured by the current meter < 2 kn

[0056] The mean square error of the flow velocity measured by the current meter < 0.2 kn

[0057] 1.2. Shallow water effect screening: For each 10 - minute time - period data set, perform the following calculation and screening:

[0058] Water depth > 80 m

[0059] Or, the distance between the bottom of the ship and the seabed > 50 m

[0060] 1.3. Stable wind and wave screening: For each 10 - minute time - period data set, perform the following calculation and screening:

[0061] The mean square error of the wind direction angle measured by the anemometer on the ship < 30 deg

[0062] The mean square error of the wave height < 0.5 m

[0063] The mean square error of the main wave direction angle of the wave < 30 deg

[0064] 2. Steps for the fusion calculation of multi - feature parameters, as Figure 1 shown, including data processing steps (i.e., mean calculation steps), water - speed conversion steps, wind - resistance increase calculation steps, still - water resistance calculation steps, and wave - resistance increase prediction steps.

[0065] 2.1. Mean calculation steps: Calculate the average value of each parameter in the ship data, meteorological data, and navigation state data respectively in the screened data set, and sort the data sets containing the average value of each parameter in chronological order. Then, successively combine several consecutive data sets in the sorted data set to form a new data set.

[0066] Specifically, after screening each data set, calculate the average value of each parameter in the ship data, meteorological data, and navigation state data respectively in the selected multiple 10 - minute data sets, such as Figure 2 the temporary 10 - minute data in the data processing shown. Then, perform screening for continuously stable navigation data, sort the data sets containing the average value of each parameter in chronological order, and then calculate the average value of the parameter averages of 3 consecutive 10 - minute data sets in the sorted data set in turn to form a new data set, that is, convert it into a 30 - minute data set and output it.

[0067] 2.2 Steps for converting the speed through water: In the new dataset, calculate the speed through water of the ship based on the ship's speed over the ground in the navigation status data and the seawater flow rate in the meteorological data. The conversion calculation of the speed through water of the ship is as follows:

[0068] a_cur_heading = abs(Cur_direction - Heading)

[0069] if a_cur_heading > 180:

[0070] a_cur_heading = 360 - a_cur_heading

[0071] Vs = Vg + V_cur * cos(a_cur_heading) (1)

[0072] Where, Vs is the speed through water of the ship, Vg is the ship's speed over the ground, Cur_direction is the predicted flow direction, V_cur is the predicted flow rate relative to the ship's navigation direction measured by the ship's current meter, Heading is the course angle, and a_cur_heading is the absolute value of the angle between the course angle and the predicted flow direction.

[0073] 2.3 Steps for calculating wind resistance: In the new dataset, calculate the actual windward area of the ship's draft based on the designed draft, the designed windward area of the draft in the ship data, and the forward draft and aft draft of the ship in the navigation status data. Calculate the wind force coefficient during the ship's navigation based on the wind direction angle in the meteorological data and the forward draft and aft draft of the ship in the navigation status data. Calculate the wind resistance based on the wind force coefficient, the wind speed in the meteorological data, and the calculated actual windward area.

[0074] Specifically, calculate the actual windward area of the ship's draft during navigation based on the designed draft, the designed windward area of the draft in the ship data, and the forward draft and aft draft of the ship in the navigation status data. The actual windward area A is calculated according to the following formula:

[0075] A = ((Ta + Tf) / 2 + 2 * (h_design_draft - (Ta + Tf) / 2)) * A_design_draft / h_design_draft (2)

[0076] In the above formula, A is the actual windward area of the draft, h_design_draft is the designed draft, A_design_draf is the designed windward area of the draft, Ta is the aft draft, and Tf is the forward draft.

[0077] Then, based on the wind direction angle in the meteorological data and the bow draft and stern draft in the navigation state data, the wind force coefficient during the actual navigation of the ship is calculated. The wind force coefficient is calculated according to the following formula:

[0078] Coefficient_wind = G(Wind_direction, Heading, Tf, Ta) (3)

[0079] In the above formula, Coefficient_wind is the wind force coefficient during the actual navigation of the ship, G(Wind_direction, Heading, Tf, Ta) is the calculation function of the wind force coefficient, which is obtained by polynomial fitting of the ship model test data measured in the ship model wind tunnel test. Wind_direction is the wind direction angle measured by the anemometer on the ship, Ta is the stern draft, and Tf is the bow draft.

[0080] Finally, based on the wind force coefficient, the wind speed in the meteorological data, and the calculated actual draft windward area, the wind resistance increase is calculated. The wind resistance increase F_wind is calculated according to the following formula:

[0081] F_wind = 0.5 * p_air * G(Wind_direction, Heading, Tf, Ta) * A * Vwind - 0.5 * p_air * G(0, 0, Tf, Ta) * A * Vg (4)

[0083] In the above formula, F_wind is the wind resistance increase, p_air is the air density, Vwind is the wind speed measured by the anemometer on the ship, A is the actual draft windward area, Wind_direction is the predicted wind direction, and G(Wind_direction, Heading, Tf, Ta) is the wind force coefficient function during the actual navigation of the ship.

[0084] 2.4. Calculation steps for still water resistance: In the new dataset, the still water resistance of the actual ship is calculated based on the bow draft, stern draft, and ship's speed over the ground in the navigation state data.

[0085] Specifically, based on the bow draft, stern draft, and ship's speed over the ground in the navigation state data, the still water resistance of the actual ship at different bow and stern drafts and different speeds is calculated. The still water resistance F_still of the actual ship is calculated according to the following formula:

[0086] F_still = F(Tf, Ta, Vg) (5)

[0087] In the above formula, F_still is the still water resistance of the actual ship, F(Tf, Ta, Vg) is the still water resistance function fitted based on the ship model experiment data, Tf is the bow draft, Ta is the stern draft, and Vg is the speed over the ground.

[0088] 2.5. Wave resistance prediction steps: In the new dataset, calculate the propeller thrust based on the wind resistance and the still water resistance of the actual ship, calculate the propeller advance speed according to the ship's speed through water, calculate the constant point based on the propeller thrust, the propeller advance speed, and the propeller diameter in the ship data, obtain the propeller advance coefficient according to the calculation function fitted from the constant point, then calculate the propeller open water efficiency under specific conditions according to the propeller advance coefficient, calculate the actual ship propulsion efficiency according to the propeller open water efficiency, the hull efficiency, the shaft transmission efficiency, and the propeller relative rotation efficiency in the ship model test data, and calculate the historical wave resistance according to the still water resistance of the actual ship, the ship's speed through water, the wind resistance, the actual ship propulsion efficiency, and the actual ship shaft power obtained based on the ship model shaft power in the ship model test data. Respectively, take the relevant characteristic parameters in the meteorological data and the navigation state data and the historical wave resistance as the training set samples, train the training set samples based on the BP neural network to obtain the wave resistance prediction model, and respectively take the relevant characteristic parameters in the meteorological data and the navigation state data as the input, and predict the wave resistance in a future time period according to the wave resistance prediction model.

[0089] Specifically, first calculate the propeller thrust based on the wind resistance and the still water resistance of the actual ship. The propeller thrust T S is calculated according to the following formula:

[0090] T S =(F_wind + F_still) / (1 - t) (6)

[0091] In the above formula, F_wind is the wind resistance, F_still is the still water resistance of the actual ship, and t is the propeller thrust deduction coefficient.

[0092] Then calculate the propeller advance speed according to the ship's speed through water. The propeller advance speed VA is calculated according to the following formula:

[0093] VA=(1 - w)*Vs (7)

[0094] In the above formula, w is the wake fraction coefficient, which can be calculated by an empirical formula or obtained through a ship model self-propulsion test, and Vs is the ship's speed through water.

[0095] Calculate the constant point K T / J 2 .

[0096] K T / J 2 =Ts*1000 / (p_s*VA 2 *D 2 ) (8)

[0097] In the above formula, K T / J 2 is a constant point, p_s is the water density, D is the propeller diameter, K T is the thrust coefficient, and J is the propeller advance coefficient.

[0098] Then, through the calculated constant point K T / J 2 and the calculation function F of the propeller advance coefficient, the propeller advance coefficient J is obtained. The propeller advance coefficient J is calculated according to the following formula:

[0099] J = F(K T _J 2 ) (9)

[0100] In the above formula, J is the propeller advance coefficient, F is the calculation function of the propeller advance coefficient, and this calculation function is obtained by plotting the curves of J and K T / J 2 on the open water characteristic curve of the propeller.

[0101] Finally, according to the propeller advance coefficient, the open water efficiency of the propeller under specific conditions is calculated. As Figure 3 shown, when the propeller advance coefficient J is exactly located on the open water test data point (J, KT / J2) of the propeller, the open water efficiency η0 of the propeller can be directly obtained. When the propeller advance coefficient J is located between any two adjacent open water test data points of the propeller, interpolation processing is performed on the open water test data of the propeller to obtain the open water efficiency η0.

[0102] Then, as Figure 4 shown, according to the open water efficiency of the propeller, as well as the hull efficiency, shafting transmission efficiency, and relative rotational efficiency of the propeller in the ship model test data, the real ship propulsion efficiency is calculated. The real ship propulsion efficiency ηDs is calculated according to the following formula:

[0103] ηDs = η0 * ηH * ηS * ηR (10)

[0104] In the above formula, η0 is the open water efficiency of the propeller, ηH is the hull efficiency, ηS is the shafting transmission efficiency, and ηR is the relative rotational efficiency of the propeller. Among them, the above data can be obtained through the open water test of the ship model or through the propeller series diagrams.

[0105] Then, according to the real ship still water resistance, ship speed through water, wind resistance increase, real ship propulsion efficiency, and real ship shaft power converted based on the ship model shaft power in the ship model test data, the historical wave resistance increase (i.e., wave resistance) of the real ship is calculated. The historical wave resistance increase is calculated according to the following formula:

[0106] F_wave_train = Pact / (Vs*ηDs)-(F_wind+F_still) (11)

[0107] In the above formula, F_wave_train is the historical wave added resistance, Pact is the real ship shaft power, Vs is the ship speed through water, F_wind is the wind added resistance, F_still is the real ship still water resistance, and ηDs is the real ship propulsion efficiency.

[0108] Finally, the wave mean period, significant wave height, main wave direction angle and course angle in the meteorological data, as well as the bow draft, stern draft and speed over the ground in the navigation state data are used as the training input data, the historical real ship wave added resistance is used as the training output data, the training input data and training output data are used as the training set samples, and the wave added resistance prediction model is trained based on the BP neural network. Then, the wave mean period, significant wave height, main wave direction angle and course angle in the meteorological data, as well as the bow draft, stern draft and speed over the ground in the navigation state data are used as the input of the wave added resistance prediction model, and the wave added resistance in a future time period is predicted according to the wave added resistance prediction model. The prediction of the wave added resistance can be calculated by the following simulation function:

[0109] F_wave = F(Tf,Ta,Vg,T_wave,H_wave,Direction_wave,Heading) (12)

[0110] In the above formula, F_wave is the predicted wave added resistance, F(Tf,Ta,Vg,T_wave,H_wave,Direction_wave,Heading) is the prediction function after BP artificial neural network training, Tf is the bow draft, Ta is the stern draft, Vg is the speed over the ground, T_wave is the predicted wave mean period, H_wave is the predicted significant wave height, Direction_wave is the predicted main wave direction angle, and Heading is the course angle.

[0111] Among them, as Figure 5 shown, the input layer of the BP neural network has a total of j node variables, the hidden layer has a total of i nodes, and the output layer has a total of k nodes. x j is the input value of the jth node, w ij represents the weight value between the ith node in the hidden layer and the jth node in the input layer, θ i represents the threshold value of the ith node in the hidden layer, φ(x) represents the activation function of the hidden layer, w ki represents the weight value between the kth node in the output layer and the ith node in the hidden layer, α k is the threshold value of the kth node in the output layer, ψ(x) represents the activation function of the output layer, and O kRepresents the node output of the final output layer. For the wave resistance increase prediction module, the number of nodes in the input layer is 7, the number of nodes in the output layer is 1, and the parameters of the intermediate layer and the solver need to be appropriately modified for the target ship.

[0112] When the BP neural network signal propagates forward, the output value O of the output layer k is calculated using the following formula:

[0113]

[0114] The backpropagation of the neural network error will correct and optimize the weights and thresholds of each layer. The error correction of the prediction result is calculated using the following formula:

[0115]

[0116] The weights and thresholds of each node after correction are calculated using the following formula:

[0117]

[0118] where j is the number of nodes in the input layer, i is the number of nodes in the hidden layer, k is the number of nodes in the output layer, x j is the input value of the j-th node, w ij represents the weight between the i-th node in the hidden layer and the j-th node in the input layer, θ i represents the threshold of the i-th node in the hidden layer, φ(x) represents the activation function of the hidden layer, w ki represents the weight between the k-th node in the output layer and the i-th node in the hidden layer, α k is the threshold of the k-th node in the output layer, ψ(x) represents the activation function of the output layer, O k represents the node output of the final output layer, T k is the actual value.

[0119] 3. Calculation steps for the main engine shaft power: Calculate the main engine shaft power based on the wind resistance increase, the predicted wave resistance increase, the calculated still water resistance of the actual ship, the ship's speed through water, and the actual ship propulsion efficiency.

[0120] Specifically, calculate the main engine shaft power based on the wind resistance increase, the predicted wave resistance increase, the still water resistance of the actual ship, the ship's speed through water, and the actual ship propulsion efficiency. The main engine shaft power PDs is calculated according to the following formula:

[0121] PDs = (F_wave + F_wind + F_still) * Vs / ηDs (19)

[0122] Among them, PDs is the main engine shaft power, F_wave is the predicted wave added resistance, F_wind is the wind added resistance, F_still is the still water resistance of the actual ship, Vs is the ship's speed through water, and ηD is the actual ship propulsion efficiency at the propeller advance coefficient J.

[0123] 4. Steps for calculating the estimated fuel consumption of the main engine: Obtain the value G(PDs) of the main engine fuel consumption rate curve function G corresponding to the main engine shaft power, and calculate the estimated fuel consumption of the main engine based on the main engine shaft power and the value of the main engine fuel consumption rate curve function. The estimated fuel consumption of the main engine is calculated according to the following formula:

[0124] Oil_pre = PDs * G(PDs) / 1000 (20)

[0125] In the above formula, Oil_pre is the estimated fuel consumption of the main engine, and G is the main engine fuel consumption rate SFOC curve function, which can be obtained from the main engine equipment manufacturer.

[0126] The present invention also relates to a ship fuel consumption estimation system, which corresponds to the above ship fuel consumption estimation method and can be understood as a system for implementing the above method. The system includes a data acquisition and screening module, a multi-feature parameter algorithm fusion calculation module, a main engine shaft power calculation module, and a main engine estimated fuel consumption calculation module connected in sequence. Among them, the multi-feature parameter algorithm fusion calculation module includes a mean calculation module, a speed-through-water conversion module, and a wind added resistance calculation module, a still water resistance calculation module, and a wave added resistance prediction module all connected to the speed-through-water conversion module. The data acquisition and screening module is connected to the mean calculation module, and the wind added resistance calculation module, the still water resistance calculation module, and the wave added resistance prediction module are all connected to the main engine shaft power calculation module. For reference Figure 1 where the data shown represents the mean calculation module, Figure 1 It can also be understood as the structural schematic diagram of the ship fuel consumption estimation system of the present invention. Specifically,

[0127] The data acquisition and screening module obtains the ship model test data measured in the ship model test, and respectively collects the ship data, meteorological data, and navigation state data during ship navigation at regular intervals, and divides all the collected data into multiple data sets at a certain time interval, and then screens each data set to screen out the data sets that meet the preset conditions;

[0128] The mean calculation module calculates the average value of each parameter in the ship data, meteorological data, and navigation state data respectively in the screened data sets, and sorts the data sets containing the average value of each parameter in the order of time, and successively combines several consecutive data sets in the sorted data sets into new data sets;

[0129] The water speed conversion module calculates the ship's speed through water in the new dataset based on the ship's speed over the ground in the navigation state data and the seawater flow rate in the meteorological data;

[0130] The wind resistance calculation module calculates the actual windward area of the draft in the new dataset based on the designed draft, the designed windward area of the draft in the ship data, and the forward draft and aft draft in the navigation state data, calculates the wind force coefficient based on the wind direction angle in the meteorological data and the forward draft and aft draft in the navigation state data, and calculates the wind resistance based on the wind force coefficient, the wind speed in the meteorological data, and the calculated actual windward area of the draft;

[0131] The calm water resistance calculation module calculates the calm water resistance of the actual ship in the new dataset based on the forward draft, the aft draft, and the ship's speed over the ground in the navigation state data;

[0132] The wave resistance prediction module calculates the propeller thrust based on the wind resistance and the calm water resistance of the actual ship in the new dataset, calculates the propeller advance speed based on the ship's speed through water, calculates the constant point based on the propeller thrust, the propeller advance speed, and the propeller diameter in the ship data, obtains the propeller advance coefficient according to the calculation function fitted by the constant point, then calculates the propeller open water efficiency under specific conditions according to the propeller advance coefficient, calculates the actual ship propulsion efficiency according to the propeller open water efficiency, the hull efficiency, the shafting transmission efficiency, and the propeller relative rotation efficiency in the ship model test data, and calculates the historical wave resistance according to the calm water resistance of the actual ship, the ship's speed through water, the wind resistance, the actual ship propulsion efficiency, and the actual ship shaft power obtained based on the ship model shaft power in the ship model test data. The relevant characteristic parameters in the meteorological data and the navigation state data and the historical wave resistance are used as training set samples respectively, and the wave resistance prediction model is trained based on the BP neural network for the training set samples. The relevant characteristic parameters in the meteorological data and the navigation state data are used as inputs respectively, and the wave resistance in a future time period is predicted according to the wave resistance prediction model;

[0133] The main engine shaft power calculation module calculates the main engine shaft power in the new dataset based on the wind resistance, the predicted wave resistance, the calculated calm water resistance of the actual ship, the ship's speed through water, and the actual ship propulsion efficiency;

[0134] The main engine estimated fuel consumption calculation module obtains the value of the main engine fuel consumption rate curve function corresponding to the main engine shaft power, and calculates the main engine estimated fuel consumption according to the main engine shaft power and the value of the main engine fuel consumption rate curve function.

[0135] Preferably, in the data acquisition and screening module, the screening of each dataset includes stable navigation screening, shallow water effect screening, and stable wind and wave screening;

[0136] And / or, the ship data includes propeller diameter, designed draft, designed draft windward area and ship wet surface area, and the meteorological data includes wind speed, wind direction, flow direction, seawater flow velocity, wave average period, significant wave height and main wave direction angle; the navigation state data includes ship bow draft, ship stern draft, rudder angle, course angle, ship frictional resistance coefficient and speed over the ground.

[0137] Preferably, the relevant characteristic parameters include bow draft, stern draft, speed over the ground, course angle, wave average period, significant wave height and main wave direction angle.

[0138] Preferably, the ship model test includes ship model self-propulsion test, ship model wind tunnel test and ship model propeller open water test.

[0139] Preferably, in the wave added resistance prediction module, calculating the propeller open water efficiency under specific conditions according to the propeller advance coefficient includes:

[0140] In the propeller open water characteristic curve graph constructed based on the propeller advance coefficient, constant points and propeller open water efficiency, when the propeller advance coefficient is located on the propeller open water experimental data points, the propeller open water efficiency is directly obtained; when the propeller advance coefficient is located between any two adjacent propeller open water experimental data points, interpolation processing is performed on the propeller open water experimental data to obtain the propeller open water efficiency.

[0141] The present invention provides an objective and scientific method and system for predicting ship fuel consumption, and is also a method and system for predicting ship fuel consumption that integrates multi-characteristic parameter algorithms. Based on ship model test data, ship data, meteorological data and navigation state data, and using a BP artificial neural network and a specific calculation method to calculate the predicted main engine fuel consumption, effectively improving the accuracy of ship fuel consumption prediction, so as to facilitate crew members to optimize ship speed and route.

[0142] It should be noted that the above specific embodiments can enable those skilled in the art to understand the present invention more comprehensively, but do not limit the present invention in any way. Therefore, although this specification has described the present invention in detail with reference to the drawings and embodiments, those skilled in the art should understand that the present invention can still be modified or equivalently replaced. In short, all technical solutions and their improvements that do not depart from the spirit and scope of the present invention should be covered by the protection scope of the patent of the present invention.

Claims

1. A method for predicting ship fuel consumption, characterized in that, It includes the following steps: Data acquisition and screening step: Obtain the ship model test data measured in the ship model test, and collect the ship data, meteorological data, and navigation status data during ship navigation at regular intervals respectively. Then divide all the collected data into multiple data sets at a certain time interval, and screen each data set to screen out the data sets that meet the preset conditions; Multi-feature parameter algorithm fusion calculation step, including mean calculation step, water speed conversion step, wind resistance increase calculation step, still water resistance calculation step, and wave resistance increase prediction step, The mean calculation step: Calculate the average value of each parameter in the ship data, meteorological data, and navigation status data respectively in the screened data set, and sort the data sets containing the average value of each parameter in chronological order. Then successively combine several consecutive data sets in the sorted data set to form a new data set; The water speed conversion step: In the new data set, calculate the ship's speed through water according to the ship's speed over the ground in the navigation status data and the seawater flow speed in the meteorological data; The wind resistance increase calculation step: In the new data set, calculate the actual windward area of the ship's draft according to the designed draft, designed draft windward area in the ship data, and the bow draft and stern draft in the navigation status data. Calculate the wind force coefficient according to the wind direction angle in the meteorological data and the bow draft and stern draft in the navigation status data. Calculate the wind resistance increase according to the wind force coefficient, the wind speed in the meteorological data, and the calculated actual windward area of the ship's draft; The still water resistance calculation step: In the new data set, calculate the actual ship's still water resistance according to the bow draft, stern draft, and ship's speed over the ground in the navigation status data; The wave resistance increase prediction step: In the new data set, calculate the propeller thrust according to the wind resistance increase and the actual ship's still water resistance, and calculate the propeller advance speed according to the ship's speed through water. Calculate the constant point according to the propeller thrust, propeller advance speed, and propeller diameter in the ship data. Obtain the propeller advance coefficient according to the calculation function fitted by the constant point. Then calculate the propeller open water efficiency under specific conditions according to the propeller advance coefficient. Calculate the actual ship's propulsion efficiency according to the propeller open water efficiency, the hull efficiency, shaft transmission efficiency, and propeller relative rotation efficiency in the ship model test data. Calculate the historical wave resistance increase according to the actual ship's still water resistance, ship's speed through water, wind resistance increase, actual ship's propulsion efficiency, and actual ship's shaft power obtained based on the ship model shaft power in the ship model test data. Respectively use the relevant characteristic parameters in the meteorological data and navigation status data and the historical wave resistance increase as training set samples, train the training set samples based on the BP neural network to obtain a wave resistance increase prediction model, and respectively use the relevant characteristic parameters in the meteorological data and navigation status data as inputs, and predict the wave resistance increase in a future period according to the wave resistance increase prediction model; Main engine shaft power calculation step: In the new data set, calculate the main engine shaft power according to the wind resistance increase, predicted wave resistance increase, calculated actual ship's still water resistance, ship's speed through water, and actual ship's propulsion efficiency; Steps for calculating the estimated fuel consumption of the main engine: Obtain the value of the main engine fuel consumption rate curve function corresponding to the main engine shaft power, and calculate the estimated fuel consumption of the main engine based on the main engine shaft power and the value of the main engine fuel consumption rate curve function.

2. The ship fuel consumption prediction method according to claim 1, wherein In the data collection and screening steps, screening each data set includes steady navigation screening, shallow water effect screening, and steady wind and wave screening. And / or, the ship data includes propeller diameter, designed draft, designed draft windward area, and ship wet surface area, the meteorological data includes wind speed, wind direction, flow direction, wind direction angle, seawater flow velocity, wave average period, significant wave height, and main wave direction angle; the navigation state data includes ship bow draft, ship stern draft, rudder angle, course angle, ship frictional resistance coefficient, and speed over the ground.

3. The ship fuel consumption prediction method according to claim 1, wherein, In the wave added resistance prediction step, the relevant characteristic parameters include bow draft, stern draft, speed over the ground, course angle, wave average period, significant wave height, and main wave direction angle.

4. The method for predicting ship fuel consumption according to claim 1, wherein The ship model test includes ship model self-propulsion test, ship model wind tunnel test, and ship model propeller open water test.

5. The ship fuel consumption prediction method according to claim 1, wherein In the wave added resistance prediction step, calculating the propeller open water efficiency under specific conditions based on the propeller advance coefficient includes: In the propeller open water characteristic curve graph constructed based on the propeller advance coefficient, constant points, and propeller open water efficiency, when the propeller advance coefficient is located at the propeller open water experimental data points, directly obtain the propeller open water efficiency; when the propeller advance coefficient is located between any two adjacent propeller open water experimental data points, perform interpolation processing on the propeller open water experimental data to obtain the propeller open water efficiency.

6. A ship fuel consumption prediction system, characterized in that, It includes a data collection and screening module, a multi-characteristic parameter algorithm fusion calculation module, a main engine shaft power calculation module, and a main engine estimated fuel consumption calculation module connected in sequence. The multi-characteristic parameter algorithm fusion calculation module includes a mean value calculation module, a speed through water conversion module, and a wind added resistance calculation module, a still water resistance calculation module, and a wave added resistance prediction module that are all connected to the speed through water conversion module. The data collection and screening module is connected to the mean value calculation module, and the wind added resistance calculation module, the still water resistance calculation module, and the wave added resistance prediction module are all connected to the main engine shaft power calculation module. The data collection and screening module obtains the ship model test data measured in the ship model test, and respectively collects the ship data, meteorological data, and navigation state data during ship navigation at regular intervals, and divides all the collected data into multiple data sets according to a certain time interval, and then screens each data set to screen out the data sets that meet the preset conditions. The mean value calculation module calculates the average value of each parameter in the ship data, meteorological data, and navigation state data respectively in the screened data sets, and sorts the data sets containing the average value of each parameter in chronological order, and successively combines several consecutive data sets in the sorted data sets into new data sets. The speed through water conversion module calculates the speed through water of the ship based on the speed over the ground of the ship in the navigation state data and the seawater flow velocity in the meteorological data in the new data set. The wind resistance increase calculation module, in the new dataset, calculates the actual windward area of the ship's draft based on the designed draft, the designed draft windward area in the ship data, and the bow draft and stern draft in the navigation state data, calculates the wind force coefficient based on the wind direction angle in the meteorological data and the bow draft and stern draft in the navigation state data, and calculates the wind resistance increase based on the wind force coefficient, the wind speed in the meteorological data, and the calculated actual windward area of the draft; The still water resistance calculation module: In the new dataset, calculates the actual ship still water resistance based on the bow draft, stern draft, and ship's speed over the ground in the navigation state data; The wave resistance prediction module, in the new dataset, calculates the propeller thrust based on the wind resistance increase and the actual ship still water resistance, calculates the propeller advance speed based on the ship's speed through water, calculates the constant point based on the propeller thrust, propeller advance speed, and propeller diameter in the ship data, obtains the propeller advance coefficient according to the calculation function fitted by the constant point, then calculates the propeller open water efficiency under specific conditions according to the propeller advance coefficient, calculates the actual ship propulsion efficiency according to the propeller open water efficiency, the hull efficiency, shaft transmission efficiency, and propeller relative rotation efficiency in the ship model test data, and calculates the historical wave resistance based on the actual ship still water resistance, ship's speed through water, wind resistance increase, actual ship propulsion efficiency, and actual ship shaft power obtained based on the ship model shaft power in the ship model test data. Respectively, the relevant characteristic parameters in the meteorological data and navigation state data and the historical wave resistance are used as training set samples, and the wave resistance prediction model is trained based on the BP neural network for the training set samples. Then, the relevant characteristic parameters in the meteorological data and navigation state data are used as inputs respectively, and the wave resistance in a future time period is predicted according to the wave resistance prediction model; The main engine shaft power calculation module, in the new dataset, calculates the main engine shaft power based on the wind resistance increase, the predicted wave resistance, the calculated actual ship still water resistance, the ship's speed through water, and the actual ship propulsion efficiency; The main engine estimated fuel consumption calculation module obtains the value of the main engine fuel consumption rate curve function corresponding to the main engine shaft power, and calculates the main engine estimated fuel consumption based on the main engine shaft power and the value of the main engine fuel consumption rate curve function.

7. The ship fuel consumption prediction system according to claim 6, characterized in that, In the data acquisition and screening module, the screening of each dataset includes steady navigation screening, shallow water effect screening, and steady wind and wave screening; And / or, the ship data includes propeller diameter, designed draft, designed draft windward area, and ship wet surface area, and the meteorological data includes wind speed, wind direction, flow direction, seawater flow velocity, wave average period, significant wave height, and main wave direction angle; the navigation state data includes ship bow draft, ship stern draft, rudder angle, course angle, ship frictional resistance coefficient, and speed over the ground.

8. The ship fuel consumption prediction system according to claim 6, wherein The relevant characteristic parameters include bow draft, stern draft, speed over the ground, course angle, wave average period, significant wave height, and main wave direction angle.

9. The ship fuel consumption prediction system according to claim 6, wherein, The ship model test includes ship model self-propulsion test, ship model wind tunnel test, and ship model propeller open water test.

10. The ship fuel consumption prediction system according to claim 6, characterized in that In the wave resistance prediction module, calculating the propeller open water efficiency under specific conditions according to the propeller advance coefficient includes: In the propeller open-water characteristic curve constructed based on the propeller advance coefficient, constant points, and propeller open-water efficiency, when the propeller advance coefficient is located at the propeller open-water experiment data points, the propeller open-water efficiency is directly obtained; when the propeller advance coefficient is located between any two adjacent propeller open-water experiment data points, interpolation processing is performed on the propeller open-water experiment data to obtain the propeller open-water efficiency.

Images