[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN116522211B - Ship oil consumption interpretable ash box estimation method, device and readable storage medium - Google Patents

Ship oil consumption interpretable ash box estimation method, device and readable storage medium Download PDF

Info

Publication number
CN116522211B
CN116522211B CN202310805635.8A CN202310805635A CN116522211B CN 116522211 B CN116522211 B CN 116522211B CN 202310805635 A CN202310805635 A CN 202310805635A CN 116522211 B CN116522211 B CN 116522211B
Authority
CN
China
Prior art keywords
ship
navigation
estimation
model
estimation model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310805635.8A
Other languages
Chinese (zh)
Other versions
CN116522211A (en
Inventor
赵宇哲
余佳昊
匡海波
周晶淼
盛尊阔
彭重秀
王楠
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dalian Maritime University
Elane Inc
Original Assignee
Dalian Maritime University
Elane Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dalian Maritime University, Elane Inc filed Critical Dalian Maritime University
Priority to CN202310805635.8A priority Critical patent/CN116522211B/en
Publication of CN116522211A publication Critical patent/CN116522211A/en
Application granted granted Critical
Publication of CN116522211B publication Critical patent/CN116522211B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10Complex mathematical operations
    • G06F17/11Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/10Pre-processing; Data cleansing
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Operations Research (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Navigation (AREA)

Abstract

The invention relates to the technical field of ship engineering, and provides a method and a device for estimating an interpretable ash box of ship oil consumption and a readable storage medium, wherein the method for estimating the interpretable ash box of the ship oil consumption comprises the following steps: acquiring a plurality of navigation parameter types in the navigation process of the ship; establishing a first estimation model between the oil consumption of the ship and a plurality of navigation parameter types; collecting historical navigation parameters corresponding to a plurality of navigation parameter types; training a preset model according to the historical navigation parameters to generate a second estimation model; and estimating the oil consumption of the ship according to the first estimation model and the second estimation model.

Description

Ship oil consumption interpretable ash box estimation method, device and readable storage medium
Technical Field
The invention relates to the technical field of ship engineering, in particular to a ship oil consumption interpretable ash box estimation method, a device and a readable storage medium.
Background
In the related art, in the estimation process of the ship oil consumption, the accuracy and the interpretability of the oil consumption estimation cannot be considered, so that the related technicians cannot trust the final estimation result, and the ship energy efficiency cannot be optimized accurately. Therefore, how to improve the accuracy of estimation of the fuel consumption of the ship and simultaneously consider the interpretability of the estimation result becomes a technical problem to be solved urgently.
Disclosure of Invention
The invention aims to at least solve the technical problem that the estimation accuracy of the ship oil consumption cannot be improved and the interpretation of the estimation result is considered in the prior art.
To this end, a first aspect of the invention provides a method for estimating a ship fuel consumption interpretable gray bin based on multi-source heterogeneous sensing data.
The second aspect of the invention provides a ship oil consumption interpretability ash bin estimation device based on multi-source heterogeneous sensing data.
The third aspect of the invention provides a ship oil consumption interpretable ash box estimation device based on multi-source heterogeneous sensing data.
A fourth aspect of the present invention provides a readable storage medium.
The first aspect of the invention provides a ship oil consumption interpretable ash bin estimation method based on multi-source heterogeneous sensing data, which comprises the following steps: acquiring a plurality of navigation parameter types in the navigation process of the ship; establishing a first estimation model between the oil consumption of the ship and a plurality of navigation parameter types; collecting historical navigation parameters corresponding to a plurality of navigation parameter types; training a preset model according to the historical navigation parameters to generate a second estimation model; and estimating the oil consumption of the ship according to the first estimation model and the second estimation model.
The method for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data can be used for estimating the fuel consumption of the ship in the process of sailing the ship so as to enable technicians to analyze the sailing process of the ship, further adjust the sailing parameters of the ship according to the estimated fuel consumption of the ship and optimize the sailing efficiency of the ship.
Specifically, first, a plurality of navigation parameter types of a ship during navigation are acquired. It will be appreciated that during actual sailing, a vessel has a number of different sailing parameter types, such as fuel temperature, fuel density, port draft, starboard draft, etc. of the vessel. Then, according to the various navigation parameter types of the ship, a relation between the oil consumption of the ship and the various navigation parameter types, namely, a first estimation model between the oil consumption of the ship and the various navigation parameter types, can be deduced. Through the first estimation model, the oil consumption of the ship can be estimated preliminarily according to various navigation parameter types generated in the navigation process of the ship. The first estimation model is established based on the mapping relation between the oil consumption of the ship and various navigation parameter types, namely, the first estimation model is calculated based on the application of a physical principle basis and a fluid dynamics law, so that the first estimation model can acquire enough priori information in the estimation process, and has good interpretation, namely, the influence of each navigation parameter type on the oil consumption of the ship can be clearly obtained.
Further, historical navigation parameters corresponding to a plurality of navigation parameter types of the ship are collected, namely, in the historical navigation process, actual navigation parameters corresponding to the plurality of navigation parameter types in a preset time period are collected. For example, during a 24 hour voyage, the fuel temperature of the ship, i.e., 288 fuel temperature histories, is obtained every 5 minutes.
And then training a preset model according to the collected historical navigation data corresponding to the navigation parameter types, namely inputting the historical navigation data into the preset model, training the preset model, and obtaining a second estimation model for estimating the oil consumption of the ship through training the historical navigation data. It can be understood that the second estimation model can directly construct the mapping relation between the navigation parameter type of the ship and the oil consumption of the ship, and the accuracy of the oil consumption estimation of the ship can be effectively improved through training of a large amount of historical navigation data.
Further, after the first estimation model and the second estimation model are obtained, the first estimation model and the second estimation model can be combined, and the estimation of the oil consumption of the ship can be achieved according to specific navigation parameters of the ship. The fuel consumption of the ship is estimated by combining the first estimation model and the second estimation model. Therefore, prior information can be provided for each navigation parameter type in the oil consumption estimation process of the ship through the first estimation model, so that the interpretability of the oil consumption estimation of the ship is guaranteed, and meanwhile, the second estimation model is obtained based on a large number of historical navigation parameters in a training mode, so that the accuracy of the oil consumption estimation of the ship can be guaranteed. That is, the fuel consumption of the ship is estimated by combining the first estimation model and the second estimation model, so that the accuracy of the fuel consumption estimation of the ship can be ensured, the interpretability of the fuel consumption estimation can be ensured, the related technicians can trust the final estimation result, and the optimization of the ship energy efficiency according to the accurate fuel consumption estimation process of the ship is facilitated for the technicians.
According to the ship oil consumption interpretable ash box estimation method based on the multi-source heterogeneous sensing data, provided by the invention, the oil consumption of the ship is estimated by establishing a first estimation model between the oil consumption of the ship and multiple navigation parameter types in the navigation process of the ship, training a preset model through a large amount of historical navigation data to obtain a second estimation model of the oil consumption of the ship, and further combining the first estimation model and the second estimation model. Therefore, the first estimation model can be utilized to provide interpretable priori knowledge based on a physical principle, so that technicians can know the influence of multiple navigation parameter types of the ship on the ship oil consumption in the ship estimation process, the interpretability of the estimation result is improved, further, the ship enterprises and maritime organization personnel are helped to deeply understand and apply the oil consumption model, and the operation energy efficiency of the ship is further improved. Meanwhile, the second estimation model is built based on a large number of historical sailing parameters and related preset models, and accuracy of estimation of oil consumption of the ship is guaranteed.
In addition, according to the method for estimating the ship oil consumption interpretable ash box based on the multi-source heterogeneous sensing data in the technical scheme provided by the invention, the method can also have the following additional technical characteristics:
In the above technical solution, further, training the preset model according to the historical navigation parameters, and before generating the second estimation model, the estimation method further includes: preprocessing the historical navigation parameters; and frequency integration is carried out on the preprocessed historical navigation parameters.
In the technical scheme, before a large number of historical navigation parameters are input into the preset model to train the second estimation model, the large number of historical navigation parameters can be preprocessed and integrated in frequency, so that the accuracy and the uniformity of the large number of historical data are guaranteed, the influence of invalid or abnormal parameters on a training result in the training process of the second estimation model is avoided, the data availability and the quality of the data for the data-driven oil consumption estimation model are improved, and the training efficiency of the preset model in the training process according to the historical navigation parameters is guaranteed.
In any of the above technical solutions, further, preprocessing the historical navigation parameters includes: deleting repeated data in the historical navigation parameters corresponding to each navigation parameter type; deleting abnormal data in the historical navigation parameters corresponding to each navigation parameter type; determining a missing value in the historical navigation parameters corresponding to each navigation parameter type according to a first preset algorithm; the missing values are supplemented into the historical voyage parameters.
In this technical solution, the preprocessing of the historical voyage parameters may include deletion of duplicate data and anomaly data and replenishment of missing values.
Specifically, the repeated data is deleted, namely the same data collected at the same time point in the collected historical navigation parameters is deleted, only one effective data is reserved, and the reduction of the model training efficiency caused by excessive repeated data is avoided.
In the process of deleting the abnormal data, firstly, the abnormal data can be identified according to a box graph algorithm, specifically, one navigation parameter type comprises a plurality of corresponding navigation parameters, and the upper limit and the lower limit of the navigation parameters are determined according to the box graph algorithmWill exceed the upper and lower limits of the box diagram +.>Is regarded as abnormal data, and is rejected. Wherein (1)>And->The method can be obtained by the following formula:Where Q3 is the lower quartile, Q1 is the upper quartile, IQR is the quartile inner distance (deviation of the upper quartile from the lower quartile), i.e., iqr=q1-Q3.
Further, the preprocessing of the historical sailing parameters further includes supplementing missing values, it can be understood that in one sailing parameter type, due to objective reasons such as actual operation conditions of the data acquisition device, there may be missing parameters at a small number of time points within a preset time period, if there is a small number of missing parameters, the imported parameters may not belong to a complete ship oil consumption characteristic segment, such parameters will seriously affect the estimation accuracy, and the linear interpolation method is used to supplement partial missing values due to factors such as equipment stability and signal transmission in the parameter acquisition process, where the formula is as follows:
Wherein y is 0 And y 1 Respectively two at t 0 And t 1 Historical navigation parameters known at the moment, y being expressed inThe historical navigation parameters missing at the moment, namely missing values.
In any of the above technical solutions, further, performing frequency integration on the preprocessed historical voyage parameters includes: determining a target time point for collecting historical navigation parameters according to the granularity of the preset time; and deleting other historical navigation parameters which are not at the target time point in the historical navigation parameters corresponding to each navigation parameter type, so that the plurality of navigation parameter types have the same acquisition frequency.
According to the technical scheme, the collection frequencies of different navigation parameter types are integrated according to the difference of the collection frequencies of the different navigation parameter types, so that the plurality of navigation parameter types have the same collection frequency, historical navigation parameters related to ship oil consumption can be extracted at the same time point, and a basis is provided for analysis of sensor data of different sampling frequencies on the same time transverse axis.
Specifically, according to the preset time granularity, the historical navigation parameters of different acquisition frequencies can be converted into acquisition frequencies corresponding to the preset time granularity, so that the frequency integration of a plurality of navigation parameter types is realized. For example, the preset time granularity may be set to 3s, so that the historical voyage parameters of each voyage parameter type are collected every 3s, and frequency integration is achieved.
In any of the above solutions, further, establishing a first estimation model according to a plurality of navigation parameter types includes: establishing an association relationship between the actual power of an engine of the ship and a plurality of navigation parameter types in the navigation process; and establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine.
In the technical scheme, in the process of establishing the first estimation model, the association relationship between the actual power of the engine of the ship and a plurality of navigation parameter types can be established first. Specifically, the association relationship is:
wherein,the ship body efficiency, the open water efficiency, the relative rotation rate and the shafting efficiency are respectively obtained through reading of a technical instruction of a host machine.
Efficiency of water supplyThe calculation can be made by the following formula:
wherein v is p For the propeller speed, n p The rotating speed of the propeller is Q, the torque of the propeller of the ship is Q, and the calculation formula of Q isT is the thrust of the propeller of the ship, and the calculation formula of T is +.>. Wherein K is T Is the thrust coefficient, K Q Is a torque coefficient; ρ is water density in kg/m 3 ;D p The diameter of the propeller is m; n is n p The unit is r ⁄ s, which is the rotation speed of the propeller. Thrust force Coefficient K T And torque coefficient K Q Can be obtained by interpolation calculation of Wageningen B series atlas.
Further, hull efficiencyThe calculation can be made by the following formula:
where t is the thrust derating coefficient and ω is the wake fraction.
The thrust derating coefficient t is defined as follows:
the definition of the companion stream score ω is as follows:
wherein v is s For the propeller speed, F p Thrust generated for the propeller; t (T) p Is the effective thrust generated by the propeller.
Further, P ME Is the actual power of the engine of the ship, P e Is the effective power in the sailing process of the ship, P e The derivation can be performed by the following procedure:
P e =R t u is provided; wherein R is t =R f +R r +R ap +R a +R aw
Wherein R is t Representing the total resistance of the ship, R f Represents friction resistance, R r R is the residual resistance ap R is the appendage resistance a R is air resistance aw For wave drag increase, U is the ship speed.
Frictional resistance R f The calculation can be made by the following formula:
wherein C is f In order to provide a coefficient of friction resistance,expressed as a surface roughness patch coefficient, ρ is the fluid mass density, U is the vessel speed, and S is the hull wet surface area. The coefficient of friction can be obtained by:
where Re is the Reynolds number, can be obtained by:
where L is the vessel length and v is the kinematic viscosity coefficient of the fluid (here water). In particular, the method comprises the steps of, The value of v is 1.5650 and 0.35.
The hull wet surface area S can be obtained by:
wherein d is draft height, B is ship-shaped width, C B Is a square coefficient. The square coefficient of the ship can be derived from the following formula:
wherein,is the volume of the drain.
Residual resistance R of ship r The calculation can be made by the following formula:
wherein F is n The Friedel number is obtained by the following formula:
appendage resistance R ap The calculation can be made by the following formula:
wherein C is ap The value of the appendage coefficient in actual calculation is 7,g as the gravitational acceleration.
Air resistance R a The calculation can be made by the following formula:
wherein C is a The specific value is 0.09, ρ a Defined as the mass density of air, is usually taken,A t Is the projection area v of the cross section of the part above the water surface of the ship a Defined as the relative speed of the ship and air.
Wave resistance increasing R aw Can be obtained by power definition calculation:
wherein,is wave height.
And finally, establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine. The preset fuel consumption rate of the engine can be obtained from the specification of the host machine.
Specifically, the first estimation model is b e = P ME ·g e Wherein b e Is the oil consumption per unit time of the ship, P ME G is the actual power of the engine e Is a preset fuel consumption rate of the engine.
In any of the above technical solutions, further, estimating the fuel consumption of the ship according to the first estimation model and the second estimation model includes: acquiring actual navigation parameters corresponding to a plurality of navigation parameter types; inputting the actual navigation parameters into a first estimation model to generate an intermediate estimation result; inputting the actual sailing parameters and the intermediate estimation results into a second estimation model to generate final estimation results.
In the technical scheme, after the first estimation model and the second estimation model are obtained, the first estimation model and the second estimation model can be combined, and the estimation of the oil consumption of the ship can be realized according to the specific navigation parameters of the ship.
Specifically, firstly, acquiring actual navigation parameters corresponding to a plurality of navigation parameter types; inputting the actual navigation parameters into a first estimation model to generate an intermediate estimation result; and finally, inputting the actual navigation parameters and the intermediate estimation result into a second estimation model to generate a final estimation result. The method comprises the steps of establishing a first estimation model and a second estimation model in a serial mode, establishing a physical balance equation between propulsion power and ship resistance through the theory of the first estimation model, determining independent variables such as relative wind speed, ship navigational speed, screw pitch, speed of relative water of a ship, ship draft height, rudder angle and ship host power, and obtaining final ship oil consumption estimation by taking the fuel consumption estimated by the first estimation model as new characteristics, wherein the fuel consumption estimated by the second estimation model is approximately regarded as a mapping relation of internal combustion oil consumption of the ship host in unit time (work of the host in unit time is fuel consumption heat value and work in unit time is instantaneous power), and the ship oil consumption is estimated by the second estimation model according to prior information given by the first estimation model and input historical navigation parameters.
Further, after the fuel consumption estimation of the ship is completed, the estimation result may be verified, specifically, the verification formula is:wherein->Is a true value;Is the average of the true values, and +.>Is the estimated value of the true value, n is the number of historical sailing parameters, R 2 Is the verification result.
In any of the above technical solutions, further, after estimating the fuel consumption of the ship according to the first estimation model and the second estimation model, the estimation method further includes: calculating a contribution value of each navigation parameter type according to the estimation result and a second preset algorithm; the contribution value is used for representing the influence degree of each navigation parameter type on the oil consumption of the ship.
In this technical solution, after the fuel consumption estimation of the ship is completed, a contribution value of each navigation parameter type in the fuel consumption estimation of the ship, that is, an influence of each navigation parameter type on the fuel consumption of the ship, may also be calculated according to a second preset algorithm. The contribution of each navigation parameter in each navigation parameter type to the fuel consumption estimation is well explained, and finally the navigation parameters utilized by the fuel consumption estimation are all interpretable.
Specifically, the contribution value of each navigation parameter can be calculated through a SHAP algorithm, wherein SHAP is a calculation method of shape value based on game theory, and each feature has a set of shape values for calculating local interpretability. Let the ith sample be x i The j-th feature of the i-th sample is x ij The model estimates the sample as y i The whole ofThe baseline (average of all sample estimates) of the model is y baseline Then:
y i= y baseline +f(x i1 )+f(x i2 )+…f(x ij );
wherein f (x) ij ) Is x ij SHAP value of (a). Intuitively, the j-th feature in the i-th sample is the final estimated value y i When the contribution value of (2) is larger than zero, the fuel consumption is indicated to have a positive effect; on the contrary, the reverse effect is provided.
The SHAP interpretation method calculates SHAP values from the joint game theory. The SHAP value of a feature value is its contribution to the output, weighting, and summing of all possible feature value combinations. First, we define the number of feature groups included in the feature set as M, the feature subset of the partial feature composition as S, and the number of feature groups included in the feature set asIs the set of all entered voyage parameters. The SHAP value of a feature j may represent the effect of the feature on model estimation after addition. To quantify this effect, a feature subset which does not contain a feature j is first entered +. >Obtaining an estimate f of the trained model s (x s ). Then adding the feature j on the basis of S to obtain an estimate of the trained model, comparing the difference between the two model estimates, traversing all possible subsets to calculate the model estimate difference before and after the feature j is added, and taking the weighted sum as the SHAP value of the feature since the effect of preserving one feature depends on other features which have been input before in the model. In the following, the specific contribution value of each ship-sensed data feature j may be calculated by the following formula:
the larger the shape value, the higher the influence of the feature on the fuel consumption of the ship. Next the expected values of the model for the feature values in the feature set of the input vessel sensor data are defined,is an estimate of the input, where +.>Is with a subset of the entered voyage parameters +.>The SHAP value for a sample feature, which is the expected value of a function of the condition, can be expressed as:
the influence of each input sensing characteristic on the fuel consumption estimation result of a certain determined sample can be clearly shown, so that the model has local interpretation capability. On this basis, the absolute value of the SHAP value of a certain feature of all samples is averaged to be used as an importance measure of the feature, namely, the global interpretation capability of the model can be formed.
According to a second aspect of the present invention, there is provided a marine fuel consumption interpretable ash bin estimation device based on multi-source heterogeneous sensing data, comprising: the device comprises an acquisition unit, a control unit and a control unit, wherein the acquisition unit is used for acquiring a plurality of navigation parameter types in the navigation process of the ship; the system comprises a building unit, a control unit and a control unit, wherein the building unit is used for building a first estimation model between the oil consumption of the ship and a plurality of navigation parameter types; the acquisition unit is used for acquiring historical navigation parameters corresponding to the plurality of navigation parameter types; the training unit is used for training the preset model according to the historical navigation parameters to generate a second estimation model; and the estimation unit is used for estimating the oil consumption of the ship according to the first estimation model and the second estimation model.
According to the ship oil consumption interpretable ash box estimation device based on the multi-source heterogeneous sensing data, provided by the invention, the oil consumption of the ship is estimated by establishing a first estimation model between the oil consumption of the ship and a plurality of navigation parameter types in the navigation process of the ship, training a preset model through a large amount of historical navigation data to obtain a second estimation model of the oil consumption of the ship, and further combining the first estimation model and the second estimation model. Therefore, the first estimation model can be utilized to provide interpretable priori knowledge based on a physical principle, so that technicians can know the influence of multiple navigation parameter types of the ship on the ship oil consumption in the ship estimation process, further help ship enterprises and maritime organization personnel to deeply understand and apply the oil consumption model, and further help to improve the running energy efficiency of the ship. Meanwhile, the second estimation model is built based on a large number of historical sailing parameters and related preset models, and accuracy of estimation of oil consumption of the ship is guaranteed.
Further, the estimation device also comprises a preprocessing unit and a frequency integration unit, wherein the preprocessing unit is used for preprocessing the historical navigation parameters; the frequency integration unit is used for frequency integration of the preprocessed historical navigation parameters.
Further, the preprocessing unit is specifically configured to delete repeated data in the historical navigation parameters corresponding to each navigation parameter type; deleting abnormal data in the historical navigation parameters corresponding to each navigation parameter type; determining a missing value in the historical navigation parameters corresponding to each navigation parameter type according to a first preset algorithm; the missing values are supplemented into the historical voyage parameters.
Further, the frequency integration unit is specifically configured to determine a target time point for acquiring the historical navigation parameter according to a preset time granularity; and deleting other historical navigation parameters which are not at the target time point in the historical navigation parameters corresponding to each navigation parameter type, so that the plurality of navigation parameter types have the same acquisition frequency.
Further, the establishing unit is specifically configured to establish a correlation between an actual power of an engine of the ship and a plurality of navigation parameter types during navigation; and establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine.
Further, the estimation unit is specifically configured to obtain actual navigation parameters corresponding to the plurality of navigation parameter types; inputting the actual navigation parameters into a first estimation model to generate an intermediate estimation result; inputting the actual sailing parameters and the intermediate estimation results into a second estimation model to generate final estimation results.
Further, the estimation device further comprises a calculation unit, wherein the calculation unit is used for calculating the contribution value of each navigation parameter type according to the estimation result and a second preset algorithm; the contribution value is used for representing the influence degree of each navigation parameter on the oil consumption of the ship.
According to a third aspect of the present invention, there is provided a marine fuel consumption interpretable ash bin estimation device based on multi-source heterogeneous sensing data, comprising: a processor and a memory storing programs or instructions executable on the processor, which when executed by the processor implement the steps of the method for estimating a fuel consumption interpretable gray tank of a ship based on multi-source heterogeneous sensing data as provided in the first aspect.
The device for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data provided by the invention comprises a memory and a processor, and further comprises a program or an instruction stored on the memory, wherein the program or the instruction can realize the steps of the method for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data in the first aspect when being executed by the processor, so that the device for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data has all the beneficial effects of the method for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data, and is not repeated herein.
According to a fourth aspect of the present invention, there is provided a readable storage medium having stored thereon a program or instructions which, when executed by a processor, implement a method for estimating a fuel consumption of a ship interpretable gray tank based on multi-source heterogeneous sensor data according to any one of the above technical solutions.
The readable storage medium provided by the invention is provided with a program or an instruction stored thereon, and when the program or the instruction is executed by a processor, the method for estimating the fuel consumption of the ship interpretable gray box based on the multi-source heterogeneous sensing data according to any one of the technical schemes can be realized, so that the readable storage medium has all the beneficial effects of the method for estimating the fuel consumption of the ship interpretable gray box based on the multi-source heterogeneous sensing data, and is not repeated herein.
Additional aspects and advantages of the invention will be set forth in part in the description which follows, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the invention will become apparent and may be better understood from the following description of embodiments taken in conjunction with the accompanying drawings in which:
fig. 1 shows a flow diagram of a method for estimating an explanatory ash bin of fuel consumption of a ship based on multi-source heterogeneous sensing data according to an embodiment of the invention;
Fig. 2 shows a flow chart of a method for estimating an explanatory ash bin of fuel consumption of a ship based on multi-source heterogeneous sensing data according to another embodiment of the invention;
fig. 3 shows a flow chart of a method for estimating an explanatory ash bin of fuel consumption of a ship based on multi-source heterogeneous sensing data according to still another embodiment of the invention;
fig. 4 shows a flow chart of a method for estimating an explanatory ash bin of fuel consumption of a ship based on multi-source heterogeneous sensing data according to still another embodiment of the invention;
fig. 5 shows a flow chart of a method for estimating an explanatory ash bin of fuel consumption of a ship based on multi-source heterogeneous sensing data according to still another embodiment of the invention;
FIG. 6 shows a block flow diagram of a method for estimating an interpretable gray tank of marine fuel consumption based on multi-source heterogeneous sensing data, according to one embodiment of the invention;
FIG. 7 shows a schematic view of a marine propulsion system according to an embodiment of the invention;
FIG. 8 shows a schematic block diagram of engine power transfer of a marine vessel according to one embodiment of the invention;
FIG. 9 is a schematic diagram showing a training process of a second estimation model in a ship fuel consumption interpretability gray box estimation method based on multi-source heterogeneous sensing data according to an embodiment of the present invention;
FIG. 10 shows a block flow diagram of a method for estimating a ship fuel consumption interpretable gray tank based on multi-source heterogeneous sensing data, according to yet another embodiment of the present invention;
FIG. 11 shows a schematic diagram of verification results of fuel consumption estimation results of a ship according to an embodiment of the present invention;
FIG. 12 illustrates a bar graph of contribution values of various navigational parameters during estimation of fuel consumption of a vessel according to one embodiment of the invention;
fig. 13 shows a block diagram of a ship fuel consumption interpretable gray tank estimation device based on multi-source heterogeneous sensing data, according to an embodiment of the present invention.
The correspondence between the reference numerals and the component names in fig. 13 is:
the ship oil consumption interpretable gray box estimation device 600 based on multi-source heterogeneous sensing data comprises a 602 acquisition unit, a 604 establishment unit, a 606 acquisition unit, a 608 training unit and a 610 estimation unit.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, without conflict, the embodiments of the present invention and features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced otherwise than as described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
The method for estimating the fuel consumption of the ship explanatory ash box based on the multi-source heterogeneous sensing data according to some embodiments of the present invention will be described with reference to fig. 1 to 13.
As shown in fig. 1, according to an embodiment of the present invention, there is provided a method for estimating an explanatory ash box of fuel consumption of a ship based on multi-source heterogeneous sensing data, including:
s102, acquiring a plurality of navigation parameter types in the navigation process of the ship;
s104, establishing a first estimation model between the oil consumption of the ship and a plurality of navigation parameter types;
s106, collecting historical navigation parameters corresponding to the navigation parameter types;
s108, training a preset model according to the historical navigation parameters to generate a second estimation model;
s110, estimating the oil consumption of the ship according to the first estimation model and the second estimation model.
The method for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data can be used for estimating the fuel consumption of the ship in the process of sailing the ship so as to enable technicians to analyze the sailing process of the ship, further adjust the sailing parameters of the ship according to the estimated fuel consumption of the ship and optimize the sailing efficiency of the ship.
Specifically, first, a plurality of navigation parameter types of a ship during navigation are acquired. It will be appreciated that during actual sailing, a vessel has a number of different sailing parameter types, such as fuel temperature, fuel density, port draft, starboard draft, etc. of the vessel. Then, according to the various navigation parameter types of the ship, a relation between the oil consumption of the ship and the various navigation parameter types, namely, a first estimation model between the oil consumption of the ship and the various navigation parameter types, can be deduced. Through the first estimation model, the oil consumption of the ship can be estimated preliminarily according to various navigation parameter types generated in the navigation process of the ship. The first estimation model is established based on the mapping relation between the oil consumption of the ship and various navigation parameter types, namely, the first estimation model is calculated based on the application of a physical principle basis and a fluid dynamics law, so that the first estimation model can acquire enough priori information in the estimation process, and has good interpretation, namely, the influence of each navigation parameter type on the oil consumption of the ship can be clearly obtained.
Further, historical navigation parameters corresponding to a plurality of navigation parameter types of the ship are collected, namely, in the historical navigation process, actual navigation parameters corresponding to the plurality of navigation parameter types in a preset time period are collected. For example, during a 24 hour voyage, the fuel temperature of the ship, i.e., 288 fuel temperature histories, is obtained every 5 minutes.
And then training a preset model according to the collected historical navigation data corresponding to the navigation parameter types, namely inputting the historical navigation data into the preset model, training the preset model, and obtaining a second estimation model for estimating the oil consumption of the ship through training the historical navigation data. It can be understood that the second estimation model can directly construct the mapping relation between the navigation parameter type of the ship and the oil consumption of the ship, and the accuracy of the oil consumption estimation of the ship can be effectively improved through training of a large amount of historical navigation data.
Further, after the first estimation model and the second estimation model are obtained, the first estimation model and the second estimation model can be combined, and the estimation of the oil consumption of the ship can be achieved according to specific navigation parameters of the ship. The fuel consumption of the ship is estimated by combining the first estimation model and the second estimation model. Therefore, prior information can be provided for each navigation parameter type in the oil consumption estimation process of the ship through the first estimation model, so that the interpretability of the oil consumption estimation of the ship is guaranteed, and meanwhile, the second estimation model is obtained based on a large number of historical navigation parameters in a training mode, so that the accuracy of the oil consumption estimation of the ship can be guaranteed. That is, the fuel consumption of the ship is estimated by combining the first estimation model and the second estimation model, so that the accuracy of the fuel consumption estimation of the ship can be ensured, the interpretability of the fuel consumption estimation can be ensured, the related technicians can trust the final estimation result, and the optimization of the ship energy efficiency according to the accurate fuel consumption estimation process of the ship is facilitated for the technicians.
Specifically, the preset model may be a Catboost model, that is, the Catboost model is trained by historical voyage data to obtain the second estimation model. Compared with the traditional gradient lifting algorithm, the Catboost model is an ordered lifting algorithm, and the algorithm can effectively resist noise points in a training set, so that gradient estimation deviation is avoided, the unavoidable gradient deviation problem in the iteration process is solved, and model generalization capability is improved. The model training efficiency can be improved, and the estimation accuracy of the second estimation model obtained through training can be guaranteed.
In particular, during the training of the second estimation model, the historical voyage parameters may be divided into a training set and a test set, wherein the former accounts for 70% of the historical voyage parameters and the latter accounts for 30% of the historical voyage parameters. The training set is used for training the second estimation model, and the testing set is used for evaluating the estimation performance of the second estimation model.
Further, in order to ensure the robustness of the result, the application adopts a K-fold cross validation method for training a second estimation model, wherein K is set to be 5, and the flow is as follows:
(1) The 70% training set was randomly split into 5 parts without repetition;
(2) Selecting 1 part of the training set as a test set, using the rest 4 parts of the training set as a training set for training a second estimation model, obtaining a model after training on the training set, testing on the test set by using the model, and storing the estimation indexes of the model;
(3) Repeating step (2) 5 times (ensuring that each subset has one chance to be a validation set);
(4) And calculating the average value of the 5 groups of test indexes as a final result of the precision evaluation of the second estimation model so as to eliminate errors caused by randomness, and taking the average value as the performance index of the second estimation model under the current 5 groups of cross validation.
And finally, verifying the estimation performance of the second estimation model on the ship oil consumption on a 30% test set by using the trained second estimation model.
According to the ship oil consumption interpretable ash box estimation method based on the multi-source heterogeneous sensing data, provided by the invention, the oil consumption of the ship is estimated by establishing a first estimation model between the oil consumption of the ship and multiple navigation parameter types in the navigation process of the ship, training a preset model through a large amount of historical navigation data to obtain a second estimation model of the oil consumption of the ship, and further combining the first estimation model and the second estimation model. Therefore, the first estimation model can be utilized to provide interpretable priori knowledge based on a physical principle, so that technicians can know the influence of multiple navigation parameter types of the ship on the ship oil consumption in the ship estimation process, further help ship enterprises and maritime organization personnel to deeply understand and apply the oil consumption model, and further help to improve the running energy efficiency of the ship. Meanwhile, the second estimation model is built based on a large number of historical sailing parameters and related preset models, and accuracy of estimation of oil consumption of the ship is guaranteed.
According to an embodiment of the present invention, as shown in fig. 2, a method for estimating an explanatory ash box of fuel consumption of a ship based on multi-source heterogeneous sensing data is provided, including:
s202, acquiring a plurality of navigation parameter types in the navigation process of the ship;
s204, establishing a first estimation model between the oil consumption of the ship and a plurality of navigation parameter types;
s206, collecting historical navigation parameters corresponding to the navigation parameter types;
s208, preprocessing the historical navigation parameters;
s210, frequency integration is carried out on the preprocessed historical navigation parameters;
s212, training a preset model according to the historical navigation parameters to generate a second estimation model;
and S214, estimating the oil consumption of the ship according to the first estimation model and the second estimation model.
In this embodiment, before a large number of historical navigation parameters are input to the preset model to train the second estimation model, the large number of historical navigation parameters can be preprocessed and integrated in frequency, so that accuracy and uniformity of a large number of historical data are guaranteed, further, influence of invalid or abnormal parameters on a training result in the training process of the second estimation model is avoided, availability and quality of data for the data-driven fuel consumption estimation model are improved, and training efficiency of the preset model in the training process according to the historical navigation parameters is guaranteed.
In the above embodiment, further, preprocessing the historical navigation parameters includes: deleting repeated data in the historical navigation parameters corresponding to each navigation parameter type; deleting abnormal data in the historical navigation parameters corresponding to each navigation parameter type; determining a missing value in the historical navigation parameters corresponding to each navigation parameter type according to a first preset algorithm; the missing values are supplemented into the historical voyage parameters.
In this embodiment, preprocessing of historical voyage parameters may include deletion of duplicate and anomaly data and replenishment of missing values.
Specifically, the repeated data is deleted, namely the same data collected at the same time point in the collected historical navigation parameters is deleted, only one effective data is reserved, and the reduction of the model training efficiency caused by excessive repeated data is avoided.
In the process of deleting the abnormal data, firstly, the abnormal data can be identified according to a box graph algorithm, specifically, one navigation parameter type comprises a plurality of corresponding navigation parameters, and the upper limit and the lower limit of the navigation parameters are determined according to the box graph algorithmWill exceed the upper and lower limits of the box diagram +. >Is regarded as abnormal data, and is rejected. Wherein (1)>And->The method can be obtained by the following formula:Where Q3 is the lower quartile, Q1 is the upper quartile, IQR is the quartile inner distance (deviation of the upper quartile from the lower quartile), i.e., iqr=q1-Q3.
Further, the preprocessing of the historical sailing parameters further includes supplementing missing values, it can be understood that in one sailing parameter type, due to objective reasons such as actual operation conditions of the data acquisition device, there may be missing parameters at a small number of time points within a preset time period, if there is a small number of missing parameters, the imported parameters may not belong to a complete ship oil consumption characteristic segment, such parameters will seriously affect the estimation accuracy, and the linear interpolation method is used to supplement partial missing values due to factors such as equipment stability and signal transmission in the parameter acquisition process, where the formula is as follows:
;/>
wherein y is 0 And y 1 Respectively two at t 0 And t 1 Historical navigation parameters known at the moment, y being expressed inThe historical navigation parameters missing at the moment, namely missing values.
In the foregoing embodiment, further, performing frequency integration on the preprocessed historical voyage parameters includes: determining a target time point for collecting historical navigation parameters according to the granularity of the preset time; and deleting other historical navigation parameters which are not at the target time point in the historical navigation parameters corresponding to each navigation parameter type, so that the plurality of navigation parameter types have the same acquisition frequency.
In this embodiment, the collection frequencies of different navigation parameter types are frequency-integrated according to the difference of the collection frequencies of the different navigation parameter types, so that a plurality of route parameter types have the same collection frequency, historical navigation parameters related to the oil consumption of the ship can be extracted at the same time point, and a basis is provided for analyzing sensor data of different sampling frequencies on the same time horizontal axis.
Specifically, according to the preset time granularity, the historical navigation parameters of different acquisition frequencies can be converted into acquisition frequencies corresponding to the preset time granularity, so that the frequency integration of a plurality of navigation parameter types is realized. For example, the preset time granularity may be set to 3s, so that the historical voyage parameters of each voyage parameter type are collected every 3s, and frequency integration is achieved.
According to an embodiment of the present invention, as shown in fig. 3, a method for estimating an explanatory ash box of fuel consumption of a ship based on multi-source heterogeneous sensing data is provided, including:
s302, acquiring a plurality of navigation parameter types in the navigation process of the ship;
s304, establishing an association relation between the actual power of an engine of the ship and a plurality of navigation parameter types in the navigation process;
S306, establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine;
s308, collecting historical navigation parameters corresponding to a plurality of navigation parameter types;
s310, training a preset model according to historical navigation parameters to generate a second estimation model;
and S312, estimating the oil consumption of the ship according to the first estimation model and the second estimation model.
In this embodiment, in the process of establishing the first estimation model, the association between the actual power of the engine of the ship and the plurality of navigation parameter types may be established first. Specifically, the association relationship is:
wherein,the ship body efficiency, the open water efficiency, the relative rotation rate and the shafting efficiency are respectively obtained through reading of a technical instruction of a host machine.
Efficiency of water supplyThe calculation can be made by the following formula:
wherein v is p For the propeller speed, n p The rotating speed of the propeller is Q, the torque of the propeller of the ship is Q, and the calculation formula of Q isT is the thrust of the propeller of the ship, and the calculation formula of T is +.>. Wherein K is T Is the thrust coefficient, K Q Is a torque coefficient; ρ is water density in kg/m 3 ;D p The diameter of the propeller is m; n is n p The unit is r ⁄ s, which is the rotation speed of the propeller. Thrust coefficient K T And torque coefficient K Q Can be obtained by interpolation calculation of Wageningen B series atlas.
Further, hull efficiencyThe calculation can be made by the following formula:
;/>
where t is the thrust derating coefficient and ω is the wake fraction.
The thrust derating coefficient t is defined as follows:
the definition of the companion stream score ω is as follows:
wherein v is s Is the propeller speed.
Further, P ME Is the actual power of the engine of the ship, P e Is the effective power in the sailing process of the ship, P e The derivation can be performed by the following procedure:
P e =R t u is provided; wherein R is t =R f +R r +R ap +R a +R aw
Wherein R is t Representing the total resistance of the ship, R f Represents friction resistance, R r R is the residual resistance ap R is the appendage resistance a R is air resistance aw The wave increases the resistance, and U is the ship speed.
Frictional resistanceThe calculation can be made by the following formula:
wherein C is f In order to provide a coefficient of friction resistance,expressed as a surface roughness patch coefficient, ρ is the fluid mass density, U is the vessel speed, and S is the hull wet surface area. The coefficient of friction can be obtained by:
where Re is the Reynolds number, can be obtained by:
where L is the vessel length and v is the kinematic viscosity coefficient of the fluid (here water). In the examplesThe value of v is 1.5650 and 0.35.
The hull wet surface area S can be obtained by:
Wherein d is draft height, B is ship-shaped width, C B Is a square coefficient. The square coefficient of the ship can be derived from the following formula:
wherein,is the volume of the drain.
Residual resistance R of ship r The calculation can be made by the following formula:
wherein F is n The Friedel number is obtained by the following formula:
;/>
appendage resistance R ap The calculation can be made by the following formula:
wherein C is ap For the appendage factor, in the examplesThe value of (2) is 7.
Air resistance R a The calculation can be made by the following formula:
wherein C is a The air resistance coefficient, which in the example is 0.09,ρ a defined as the mass density of air, is usually takenA t Is the projection area of the cross section of the part above the water surface of the ship,v a defined as the relative speed of the ship and air.
Wave resistance increasing R aw Can be obtained by power definition calculation:
wherein,is wave height.
And finally, establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine. The preset fuel consumption rate of the engine can be obtained from the specification of the host machine.
Specifically, the first estimation model is b e = P ME ·g e Wherein b e Is the oil consumption per unit time of the ship, P ME G is the actual power of the engine e Is a preset fuel consumption rate of the engine.
Further, after the first estimation model is built, estimation performance of the first estimation model may be further estimated, specifically, a Root Mean Square Error (RMSE) and an average absolute error (MAE) between the true value and the estimated value may be estimated, which is defined as follows:
Wherein the method comprises the steps ofIs the estimated true value; but->Is an estimate of the true value, n is the number of historical voyage parameters.
According to an embodiment of the present invention, as shown in fig. 4, a method for estimating an explanatory ash box of fuel consumption of a ship based on multi-source heterogeneous sensing data is provided, including:
s402, acquiring a plurality of navigation parameter types in the navigation process of the ship;
s404, establishing an association relationship between the actual power of an engine of the ship and a plurality of navigation parameter types in the navigation process;
s406, establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine;
s408, collecting historical navigation parameters corresponding to a plurality of navigation parameter types;
s410, training a preset model according to historical navigation parameters to generate a second estimation model;
s412, acquiring actual navigation parameters corresponding to the plurality of navigation parameter types;
s414, inputting the actual sailing parameters into a first estimation model to generate an intermediate estimation result;
s416, inputting the actual sailing parameters and the intermediate estimation results into a second estimation model to generate final estimation results.
In this embodiment, after the first estimation model and the second estimation model are obtained, the first estimation model and the second estimation model may be combined, and estimation of the fuel consumption of the ship may be achieved according to specific sailing parameters of the ship.
Specifically, firstly, acquiring actual navigation parameters corresponding to a plurality of navigation parameter types; inputting the actual navigation parameters into a first estimation model to generate an intermediate estimation result; and finally, inputting the actual navigation parameters and the intermediate estimation result into a second estimation model to generate a final estimation result. The method comprises the steps of establishing a first estimation model and a second estimation model in a serial mode, establishing a physical balance equation between propulsion power and ship resistance through the theory of the first estimation model, determining independent variables such as relative wind speed, ship navigational speed, screw pitch, speed of relative water of a ship, ship draft, rudder angle and power of a ship host machine as an objective function, and obtaining final ship oil consumption estimation by taking the fuel consumption estimated by the first estimation model as new characteristics, wherein the fuel consumption estimated by the second estimation model is approximately regarded as a mapping relation of the fuel consumption of the ship host machine in unit time (work of the host machine in unit time is a fuel consumption heat value and work in unit time is instantaneous power).
Further, after the fuel consumption estimation of the ship is completed, the estimation result may be verified, specifically, the verification formula is:wherein->Is a true value;Is the average of the true values, and +.>Is the estimated value of the true value, n is the number of historical sailing parameters, R 2 Is the verification result.
According to an embodiment of the present invention, as shown in fig. 5, a method for estimating an explanatory ash box of fuel consumption of a ship based on multi-source heterogeneous sensing data is provided, including:
s502, acquiring a plurality of navigation parameter types in the navigation process of the ship;
s504, establishing an association relation between the actual power of an engine of the ship and a plurality of navigation parameter types in the navigation process;
s506, establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine;
s508, collecting historical navigation parameters corresponding to a plurality of navigation parameter types;
s510, training a preset model according to the historical navigation parameters to generate a second estimation model;
s512, acquiring actual navigation parameters corresponding to the plurality of navigation parameter types;
s514, inputting the actual sailing parameters into a first estimation model to generate an intermediate estimation result;
s516, inputting the actual sailing parameters and the intermediate estimation result into a second estimation model to generate a final estimation result;
And S518, calculating the contribution value of each navigation parameter type according to the estimation result and a second preset algorithm.
In this embodiment, after the fuel consumption of the ship is estimated, a contribution value of each navigation parameter type in the fuel consumption estimation of the ship, that is, an influence of each navigation parameter type on the fuel consumption of the ship, may also be calculated according to a second preset algorithm. The contribution of each voyage parameter type to the fuel consumption estimation is well explained, and finally the voyage parameters utilized by the estimation of the fuel consumption are all provided with interpretability.
Specifically, the contribution value of each navigation parameter can be calculated through a SHAP algorithm, wherein SHAP is a calculation method of shape value based on game theory, and each feature has a set of shape values for calculating local interpretability. Let the ith sample be x i The j-th feature of the i-th sample is x ij The model estimates the sample as y i The baseline (average of all sample estimates) of the entire model is y baseline Then:
y i= y baseline +f(x i1 )+f(x i2 )+…f(x ij );
wherein f (x) ij ) Is x ij SHAP value of (a). Intuitively, the j-th feature in the i-th sample is the final estimated value y i When the contribution value of (2) is larger than zero, the fuel consumption is indicated to have a positive effect; on the contrary, the reverse effect is provided.
The SHAP interpretation method calculates SHAP values from the joint game theory. The SHAP value of a feature value is its contribution to the output, weighting, and summing of all possible feature value combinations. First, we define the number of feature groups included in the feature set as M, the feature subset of the partial feature composition as S, and the number of feature groups included in the feature set asIs the set of all entered voyage parameters. The SHAP value of a feature j may represent the effect of the feature on model estimation after addition. To quantify this effect, a feature subset which does not contain a feature j is first entered +.>Obtaining an estimate f of the trained model s (x s ). Then adding the feature j on the basis of S to obtain an estimate of the trained model, comparing the difference between the two model estimates, traversing all possible subsets to calculate the model estimate difference before and after the feature j is added, and taking the weighted sum as the SHAP value of the feature since the effect of preserving one feature depends on other features which have been input before in the model. In the following, the specific contribution value of each ship-sensed data feature j may be calculated by the following formula:
the larger the shape value, the higher the influence of the feature on the fuel consumption of the ship. Next, expected values of the model for the characteristic values in the input ship sensing data characteristic set are defined Wherein the functionThe SHAP value for the sample feature can be expressed as:
the influence of each input sensing characteristic on the fuel consumption estimation result of a certain determined sample can be clearly shown, so that the model has local interpretation capability. On this basis, the absolute value of the SHAP value of a certain feature of all samples is averaged to be used as an importance measure of the feature, namely, the global interpretation capability of the model can be formed.
In a specific implementation process, as shown in fig. 6, first, taking a passenger ship as an example, various high-precision sensor composition sensing systems are installed at different positions of the passenger ship, external environment information and ship self information, namely a plurality of navigation parameter types, encountered in the navigation process of the ship and acquired by the sensors in a preset time period are collected by the sensors, and the power system composition of the system is shown in fig. 7, wherein the characteristics include 19 characteristics such as fuel temperature, fuel density, fuel volume flow, port water level and the like.
TABLE 1
Further, the historical navigation parameters corresponding to the navigation parameter types are collected, and the historical navigation parameters are preprocessed. Specifically, the data preprocessing includes the elimination of duplicate data and abnormal data, and the supplementation of three parts of missing data; firstly, taking the time parameter of the sensor as a standard for measuring whether the sensor is repeated, transcoding the sensor in a standard time format is needed, and if a plurality of pieces of same data appear at the same time point, the repeated data can be considered to appear and are removed.
Secondly, identifying abnormal data by a box graph method to exceedUpper and lower limits of box-shaped diagramIs regarded as abnormal data and is eliminated, < ->And->The method can be obtained by the following formula:Where Q3 is the lower quartile, Q1 is the upper quartile, IQR is the quartile inner distance (deviation of the upper quartile from the lower quartile), i.e., iqr=q1-Q3.
Further, a linear interpolation method is used for supplementing partial missing values caused by factors such as equipment stability and signal transmission in the parameter acquisition process, and the formula is as follows:
wherein y is 0 And y 1 Respectively two at t 0 And t 1 Historical navigation parameters known at the moment, y being expressed inThe historical navigation parameters missing at the moment, namely missing values.
Further, a white-box fuel consumption estimation model (i.e., the first estimation model in the above embodiment) is established.
It will be appreciated that rapidity is one of the important hydrodynamic properties of a ship. Marine rapidity refers to the ability of a marine vessel to absorb energy from a propeller to develop thrust to overcome drag to maintain a speed of travel, and includes both aspects of marine vessel drag and marine vessel propulsion. The ship white-box fuel consumption estimation model is a result of calculating the resistance encountered by the ship from different sources based on the application of physical principle basis and fluid dynamics law. By modeling the total drag condition, the corresponding fuel consumption required to drive the ship at a particular speed can be calculated.
Firstly, the ship is subjected to the resistance action of two mediums, namely water and air, in the actual sailing process, so that the total resistance of the ship can be divided into the water resistance of an underwater part and the air resistance of an above-water part.
The water resistance is further divided into two parts, namely, the hydrostatic resistance and the wave resistance increase which is suffered by the sailing, wherein the hydrostatic resistance comprises the bare hull resistance and the appendage resistance, and the appendage resistance refers to the additional resistance brought by the structure outside the bare hull in the flow field. The bare hull resistance can be classified into frictional resistance, viscous-pressure resistance, and wave-making resistance from the mechanism of resistance generation, where the viscous-pressure resistance and the wave-making resistance are collectively referred to as residual resistance. The resistance above the water surface mainly comes from the resistance of air to the ship superstructure, and the magnitude of the air resistance is related to the shape and relative wind speed of the ship water part. In addition, when the ship sails in the stormy waves, the resistance increased by the wave action is the wave resistance increase.
The power from the main engine (i.e. the engine of the ship) is first transferred to the propeller via the propeller shaft system of the ship, and the propeller obtains less power than the main engine of the ship due to the intermediate power transfer losses. In the power system, a host is an energy source, the power of the host is output through consuming fuel, the power of the ship host is transmitted to a propeller through a transmission shaft system, and the power obtained by the propeller can be influenced by shafting efficiency between the ship host and the propeller; the propeller generates thrust by rotating the obtained power, wherein the ratio of the propulsion power transmitted by the propeller to the obtained power is the product of the relative rotation efficiency and the water-opening efficiency of the propeller; the propulsion power is further transmitted to the ship body, and the ship body obtains effective power finally, wherein the ratio of the effective power obtained by the ship body to the propulsion power transmitted by the propeller is the ship body efficiency. The effective power of the ship body is finally converted into power for overcoming the resistance of the ship body, and the ship is pushed to stably sail at a certain sailing speed by overcoming the resistance of the ship, as shown in fig. 8.
According to the power transfer relation of the main engine and the transfer efficiency among the parts, the main engine power can be deduced from the effective power, and the oil consumption of the ship in unit time can be obtained. Therefore, based on the energy conservation theory, the ship stress analysis and the transmission efficiency, the built ship fuel consumption white box model is shown as the formula: b e = P ME ·g e Wherein b e Is the oil consumption per unit time of the ship, P ME G is the actual power of the engine e The preset fuel consumption rate of the engine can be derived by the derivation process in the above-described embodiment.
Further, after the first estimation model is built, estimation performance of the first estimation model may be further estimated, specifically, a Root Mean Square Error (RMSE) and an average absolute error (MAE) between the true value and the estimated value may be estimated, which is defined as follows:
wherein the method comprises the steps ofIs the estimated true value; but->Is an estimate of the true value, n is the number of historical voyage parameters.
The estimation results are shown in table 2:
TABLE 2
Further, a black tank fuel consumption estimation model (i.e., the second estimation model in the above embodiment) is established.
Specifically, the ship black box fuel consumption estimation model is a mapping relationship of input features constructed based on a machine-learned Catboost estimation algorithm and ship fuel consumption, and the Catboost algorithm can process a wider range of ship fuel consumption input features without any prior knowledge about the modeled system. The fuel consumption of the ship is estimated by the input features of the sensor. The model development environment is developed under a Windows 11 professional operating system and a Catboost1.0.6 library by using Python 3.7, and the hardware is configured as a 12th Gen Intel (R) Core (TM) i9-12900K 3.20 GHz,64 bit operating system. The main parameter set of the model is as follows: itemerations 500; loss Function; logoss; learning Rate 0.03; max Depth 6.
Before the black box model is estimated, the model needs to be trained with a portion of data so that the model learns the mapping between fuel consumption and input features. Thus, the dataset is divided into a training set and a testing set, wherein the former accounts for 70% of the original data and the latter accounts for 30% of the original data. The training set is used for training the black box model, and the testing set is used for evaluating the estimated performance of the model.
Further, in order to ensure the robustness of the result, the application adopts a K-fold cross validation method for training the second estimation model, wherein K is set to 5, as shown in fig. 9, and the flow is as follows:
(1) The 70% training set was randomly split into 5 parts without repetition;
(2) Selecting 1 part of the training set as a test set, using the rest 4 parts of the training set as a training set for training a second estimation model, obtaining a model after training on the training set, testing on the test set by using the model, and storing the estimation indexes of the model;
(3) Repeating step (2) 5 times (ensuring that each subset has one chance to be a validation set);
(4) And calculating the average value of the 5 groups of test indexes as a final result of the precision evaluation of the second estimation model so as to eliminate errors caused by randomness, and taking the average value as the performance index of the second estimation model under the current 5 groups of cross validation.
And finally, verifying the estimation performance of the second estimation model on the ship oil consumption on a 30% test set by using the trained second estimation model.
Wherein, the verification result is shown in table 3:
TABLE 3 Table 3
Then, the ash bin fuel consumption estimation model of the SHAP framework (i.e., the estimation model after combining the first estimation model and the second estimation model) is fused.
The ship ash box fuel consumption estimation model is a combined modeling method of establishing a white box model and a black box in a serial mode, a physical balance equation between propulsion power and ship resistance is established through a white box theory, independent variables such as relative wind speed, ship navigational speed, screw pitch, ship relative water speed, ship draft height, rudder angle and ship host power are determined, the ship ash box fuel consumption estimation model can be approximately regarded as a mapping relation of internal combustion fuel consumption (work of the host machine in unit time is fuel consumption heat value, work of the host machine in unit time is instantaneous power) in unit time, the fuel consumption estimated by the white box is taken as new characteristics to be input into the black box model, the black box model estimates ship fuel consumption according to priori information given by the white box model and input fuel consumption characteristic data, and final ship fuel consumption estimation is obtained, as shown in fig. 10, and estimation results are shown in table 4:
TABLE 4 Table 4
Further, after the fuel consumption estimation of the ship is completed, the estimation result may be verified, specifically, the verification formula is:wherein->Is the estimated true value; but->Is an estimate of the true value, n is the number of historical voyage parameters.
R of estimated value and true value 2 The fitting result is shown in fig. 11, the fitting result of the gray box model is very concentrated, a remarkable linear relation is presented, and a stable estimation result can be provided.
Finally, the contribution value of each navigation parameter can be calculated through a SHAP algorithm, wherein SHAP is a calculation method of shape value based on game theory, and each feature has a group of shape values for calculating local interpretability. In particular the contribution value may be calculated by the following formula:
wherein,is the number of entered voyage parameters, +.>Is the set of all entered voyage parameters, +.>Is a set of non-zero feature indices.Is an estimate of the input, where +.>Is with a subset of the entered voyage parameters +.>The larger the shape value, which is the expected value of the function of the condition, the larger the contribution of the feature. The importance of the 19 input features to fuel consumption is shown in fig. 12, so that the ship ash box fuel consumption estimation model has an interpretability.
According to a second aspect of the present invention, as shown in fig. 13, there is provided a ship fuel consumption interpretable ash box estimation device 600 based on multi-source heterogeneous sensing data, including: an obtaining unit 602, configured to obtain a plurality of navigation parameter types during navigation of a ship; a building unit 604, configured to build a first estimation model between the fuel consumption of the ship and a plurality of navigation parameter types; the collection unit 606 is configured to collect historical navigation parameters corresponding to the plurality of navigation parameter types; the training unit 608 is configured to train the preset model according to the historical navigation parameters, and generate a second estimation model; and the estimation unit 610 is used for estimating the oil consumption of the ship according to the first estimation model and the second estimation model.
According to the ship oil consumption interpretable ash box estimation device 600 based on the multi-source heterogeneous sensing data, a first estimation model between ship oil consumption and a plurality of navigation parameter types in a ship navigation process is established, a second estimation model of ship oil consumption is obtained by training a preset model through a large amount of historical navigation data, and the first estimation model and the second estimation model are further combined to realize the estimation of the ship oil consumption. Therefore, the first estimation model can be utilized to provide interpretable priori knowledge based on a physical principle, so that technicians can know the influence of multiple navigation parameter types of the ship on the ship oil consumption in the ship estimation process, further help ship enterprises and maritime organization personnel to deeply understand and apply the oil consumption model, and further help to improve the running energy efficiency of the ship. Meanwhile, the second estimation model is built based on a large number of historical sailing parameters and related preset models, and accuracy of estimation of oil consumption of the ship is guaranteed.
Further, the estimation device also comprises a preprocessing unit and a frequency integration unit, wherein the preprocessing unit is used for preprocessing the historical navigation parameters; the frequency integration unit is used for frequency integration of the preprocessed historical navigation parameters.
Further, the preprocessing unit is specifically configured to delete repeated data in the historical navigation parameters corresponding to each navigation parameter type; deleting abnormal data in the historical navigation parameters corresponding to each navigation parameter type; determining a missing value in the historical navigation parameters corresponding to each navigation parameter type according to a first preset algorithm; the missing values are supplemented into the historical voyage parameters.
Further, the frequency integration unit is specifically configured to determine a target time point for acquiring the historical navigation parameter according to a preset time granularity; and deleting other historical navigation parameters which are not at the target time point in the historical navigation parameters corresponding to each navigation parameter type, so that the plurality of navigation parameter types have the same acquisition frequency.
Further, the establishing unit 604 is specifically configured to establish an association relationship between the actual power of the engine of the ship and the plurality of navigation parameter types during navigation; and establishing a first estimation model according to the association relation and the preset fuel consumption rate of the engine.
Further, the estimation unit 610 is specifically configured to obtain actual navigation parameters corresponding to the plurality of navigation parameter types; inputting the actual navigation parameters into a first estimation model to generate an intermediate estimation result; inputting the actual sailing parameters and the intermediate estimation results into a second estimation model to generate final estimation results.
Further, the estimation device further comprises a calculation unit, wherein the calculation unit is used for calculating the contribution value of each navigation parameter type according to the estimation result and a second preset algorithm; the contribution value is used for representing the influence degree of each navigation parameter on the oil consumption of the ship.
According to a third aspect of the present invention, there is provided a marine fuel consumption interpretable ash bin estimation device based on multi-source heterogeneous sensing data, comprising: a processor and a memory storing programs or instructions executable on the processor, which when executed by the processor implement the steps of the method for estimating a fuel consumption interpretable gray tank of a ship based on multi-source heterogeneous sensing data as provided in the first aspect.
The device for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data provided by the invention comprises a memory and a processor, and further comprises a program or an instruction stored on the memory, wherein the program or the instruction can realize the steps of the method for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data in the first aspect when being executed by the processor, so that the device for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data has all the beneficial effects of the method for estimating the fuel consumption interpretable ash box of the ship based on the multi-source heterogeneous sensing data, and is not repeated herein.
According to a fourth aspect of the present invention, there is provided a readable storage medium having stored thereon a program or instructions which, when executed by a processor, implement a method for estimating a fuel consumption of a ship interpretable gray tank based on multi-source heterogeneous sensor data according to any one of the above technical solutions.
The readable storage medium provided by the invention is provided with a program or an instruction stored thereon, and when the program or the instruction is executed by a processor, the method for estimating the fuel consumption of the ship interpretable gray box based on the multi-source heterogeneous sensing data according to any one of the technical schemes can be realized, so that the readable storage medium has all the beneficial effects of the method for estimating the fuel consumption of the ship interpretable gray box based on the multi-source heterogeneous sensing data, and is not repeated herein.
In the description of the present invention, the term "plurality" means two or more, unless explicitly defined otherwise, the orientation or positional relationship indicated by the terms "upper", "lower", etc. are based on the orientation or positional relationship shown in the drawings, merely for convenience of description of the present invention and to simplify the description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present invention; the terms "coupled," "mounted," "secured," and the like are to be construed broadly, and may be fixedly coupled, detachably coupled, or integrally connected, for example; can be directly connected or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.
In the description of the present specification, the terms "one embodiment," "some embodiments," "particular embodiments," and the like, mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (6)

1. The ship oil consumption interpretable ash box estimation method based on the multi-source heterogeneous sensing data is characterized by comprising the following steps of:
acquiring a plurality of navigation parameter types in the navigation process of the ship;
establishing a first estimation model between the oil consumption of the ship and the plurality of navigation parameter types;
Collecting historical navigation parameters corresponding to the navigation parameter types;
training a preset model according to the historical navigation parameters to generate a second estimation model;
estimating the oil consumption of the ship according to the first estimation model and the second estimation model;
establishing a first estimation model according to the navigation parameter types, wherein the first estimation model comprises the following steps:
establishing an association relationship between the actual power of the engine of the ship and the plurality of navigation parameter types during navigation;
establishing the first estimation model according to the association relation and the preset fuel consumption rate of the engine;
the first estimation model can acquire priori information in the estimation process;
before training the preset model according to the historical navigation parameters and generating the second estimation model, the ship oil consumption interpretable ash box estimation method based on the multi-source heterogeneous sensing data further comprises the following steps:
preprocessing the historical navigation parameters;
frequency integration is carried out on the preprocessed historical navigation parameters;
the preprocessing of the historical navigation parameters comprises:
deleting repeated data in the historical navigation parameters corresponding to each navigation parameter type;
Deleting abnormal data in the historical navigation parameters corresponding to each navigation parameter type;
determining a missing value in the historical navigation parameters corresponding to each navigation parameter type according to a first preset algorithm;
supplementing the missing value to the historical voyage parameter;
the estimating the fuel consumption of the ship according to the first estimation model and the second estimation model comprises the following steps:
acquiring actual navigation parameters corresponding to the plurality of navigation parameter types;
inputting the actual navigation parameters into the first estimation model to generate an intermediate estimation result;
inputting the actual navigation parameters and the intermediate estimation results into the second estimation model to generate final estimation results;
the association relation is as follows:
wherein eta HPRS Respectively the hull efficiency, the open water efficiency, the relative rotation rate and the shafting efficiency, P ME Is the actual power of the engine of the ship, P e Is effective power in the sailing process of the ship;
the preset model may be a Catboost model, which is an ordered lifting algorithm.
2. The method for estimating the fuel consumption interpretable gray box of the ship based on the multi-source heterogeneous sensing data according to claim 1, wherein the frequency integration of the preprocessed historical voyage parameters comprises the following steps:
Determining a target time point for collecting the historical navigation parameters according to the preset time granularity;
deleting other historical navigation parameters which are not located at the target time point in the historical navigation parameters corresponding to each navigation parameter type, so that a plurality of navigation parameter types have the same acquisition frequency.
3. The method for estimating a fuel consumption interpretable gray box of a ship based on multi-source heterogeneous sensing data according to claim 1 or 2, wherein after the estimating of the fuel consumption of the ship according to the first estimation model and the second estimation model, the method for estimating a fuel consumption interpretable gray box of a ship based on multi-source heterogeneous sensing data further comprises:
calculating the contribution value of each navigation parameter type according to the estimation result and a second preset algorithm;
the contribution value is used for representing the influence degree of each navigation parameter type on the oil consumption of the ship.
4. The utility model provides a ship oil consumption interpretability ash bin estimation device based on multisource heterogeneous sensing data which characterized in that includes:
the device comprises an acquisition unit, a control unit and a control unit, wherein the acquisition unit is used for acquiring a plurality of navigation parameter types in the navigation process of the ship;
the establishing unit is used for establishing a first estimation model between the oil consumption of the ship and the plurality of navigation parameter types;
The acquisition unit is used for acquiring historical navigation parameters corresponding to the plurality of navigation parameter types;
the training unit is used for training a preset model according to the historical navigation parameters to generate a second estimation model;
the estimating unit is used for estimating the oil consumption of the ship according to the first estimating model and the second estimating model;
the establishing unit is specifically used for establishing an association relation between the actual power of the engine of the ship and the plurality of navigation parameter types in the navigation process; establishing the first estimation model according to the association relation and the preset fuel consumption rate of the engine;
the first estimation model can acquire priori information in the estimation process;
the preprocessing unit is used for preprocessing the historical navigation parameters;
the frequency integration unit is used for frequency integration of the preprocessed historical navigation parameters;
the preprocessing unit is specifically used for deleting repeated data in the historical navigation parameters corresponding to each navigation parameter type; deleting abnormal data in the historical navigation parameters corresponding to each navigation parameter type; determining a missing value in the historical navigation parameters corresponding to each navigation parameter type according to a first preset algorithm; supplementing the missing value to the historical voyage parameter;
The estimation unit is specifically configured to obtain actual navigation parameters corresponding to the multiple navigation parameter types; inputting the actual navigation parameters into the first estimation model to generate an intermediate estimation result; inputting the actual navigation parameters and the intermediate estimation results into the second estimation model to generate final estimation results;
the association relation is as follows:
wherein eta HPRS Respectively the hull efficiency, the open water efficiency, the relative rotation rate and the shafting efficiency, P ME Is the actual power of the engine of the ship, P e Is effective power in the sailing process of the ship;
the preset model may be a Catboost model, which is an ordered lifting algorithm.
5. The utility model provides a ship oil consumption interpretability ash bin estimation device based on multisource heterogeneous sensing data which characterized in that includes:
a processor and a memory storing programs or instructions executable on the processor, which when executed by the processor implement the steps of the method for estimating a fuel consumption of a vessel based on multisource heterogeneous sensor data according to any one of claims 1 to 3.
6. A readable storage medium, characterized in that it has stored thereon a program or instructions, which when executed by a processor, realizes the steps of the method for estimating the fuel consumption of a ship interpretable gray tank based on multisource heterogeneous sensor data according to any one of claims 1 to 3.
CN202310805635.8A 2023-07-03 2023-07-03 Ship oil consumption interpretable ash box estimation method, device and readable storage medium Active CN116522211B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310805635.8A CN116522211B (en) 2023-07-03 2023-07-03 Ship oil consumption interpretable ash box estimation method, device and readable storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310805635.8A CN116522211B (en) 2023-07-03 2023-07-03 Ship oil consumption interpretable ash box estimation method, device and readable storage medium

Publications (2)

Publication Number Publication Date
CN116522211A CN116522211A (en) 2023-08-01
CN116522211B true CN116522211B (en) 2024-02-02

Family

ID=87401531

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310805635.8A Active CN116522211B (en) 2023-07-03 2023-07-03 Ship oil consumption interpretable ash box estimation method, device and readable storage medium

Country Status (1)

Country Link
CN (1) CN116522211B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210027225A1 (en) * 2019-07-24 2021-01-28 Marine Technologies LLC System and method for optimizing fuel usage of a marine vessel
CN112446096A (en) * 2020-11-26 2021-03-05 上海船舶运输科学研究所 Ship host engine oil consumption dynamic estimation method based on big navigation data
CN115660137A (en) * 2022-09-07 2023-01-31 中远海运科技股份有限公司 Method for accurately estimating wind wave navigation energy consumption of ship

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210027225A1 (en) * 2019-07-24 2021-01-28 Marine Technologies LLC System and method for optimizing fuel usage of a marine vessel
CN112446096A (en) * 2020-11-26 2021-03-05 上海船舶运输科学研究所 Ship host engine oil consumption dynamic estimation method based on big navigation data
CN115660137A (en) * 2022-09-07 2023-01-31 中远海运科技股份有限公司 Method for accurately estimating wind wave navigation energy consumption of ship

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
"High-Dimensional, Small-Sample Product Quality Prediction Method Based on MIC-Stacking Ensemble Learning";Yu Jiahao 等;《Applied Sciences》;全文 *
"基于油耗预测的船舶航线优化研究";冒志豪;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;全文 *
"基于航行数据的船舶航行油耗模型建立方法";冯亮 等;《上海船舶运输科学研究所学报》;第43卷(第1期);全文 *
"航运资产整合下海运网络的航线、路径和船舶的集成优化";赵宇哲 等;《系统工程理论与实践》;第38卷(第8期);全文 *

Also Published As

Publication number Publication date
CN116522211A (en) 2023-08-01

Similar Documents

Publication Publication Date Title
EP3330171B1 (en) Apparatus for predicting a power consumption of a maritime vessel
CN108431708B (en) Diagnostic system and method for estimating operational efficiency of a marine vessel
CN113033073A (en) Unmanned ship energy efficiency digital twinning method and system based on data driving
JP2015526778A (en) And system for predicting ship performance
CN115342814B (en) Unmanned ship positioning method based on multi-sensor data fusion
Cao et al. Digital twin real time monitoring method of turbine blade performance based on numerical simulation
CN112088122B (en) Method and system for evaluating actual sea area propulsion performance of ship and computer readable storage medium
CN117742346A (en) Real-time navigation state analysis and energy consumption control method for ship
CN116522211B (en) Ship oil consumption interpretable ash box estimation method, device and readable storage medium
JP6846896B2 (en) Analysis of ship propulsion performance
KR101863747B1 (en) Fuel consumption analysis method for standard vessel operation condition
Pedersen et al. Modeling of ship propulsion performance
CN114065639B (en) Closed-loop real-time inversion method for construction parameters of dredger
CN114462305B (en) Method for optimizing long-endurance energy consumption and estimating service life of Argo buoy
JP2023051248A (en) Performance evaluation method, performance evaluation program, performance evaluation system by actual monitoring analysis of vessel
Han et al. A RP image-based health indicator construction method for wind turbine RUL prediction
Wang et al. A SVM-based Sensitivity Analysis Approach for Data-Driven Modeling of Ship Motion
CN114638152B (en) Deep sea Argo profile buoy energy management method based on HGP-MPC
Wahl Prediction of fuel consumption of a ship in transit using machine learning
Song et al. Uncertainty-Aware Acoustic Localization and Mapping for Underwater Robots
TWI792418B (en) Method for predicting fuel consumption of ship
Zalachoris Evaluation of the efficiency of a Mewis duct by performance monitoring
CN118655334A (en) Ocean current flow speed forecasting method and device and electronic equipment
da Paixão et al. Comparison of volume and centroid uncertainty of a cylindrical segment by two different methods
Ponte Machine Learning-accelerated computational fluid dynamics for naval ship resistance prediction

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant