CN108052707B - Ship navigation condition division method based on cluster analysis - Google Patents
Ship navigation condition division method based on cluster analysis Download PDFInfo
- Publication number
- CN108052707B CN108052707B CN201711214624.3A CN201711214624A CN108052707B CN 108052707 B CN108052707 B CN 108052707B CN 201711214624 A CN201711214624 A CN 201711214624A CN 108052707 B CN108052707 B CN 108052707B
- Authority
- CN
- China
- Prior art keywords
- data
- ship
- cluster analysis
- condition
- wind speed
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Probability & Statistics with Applications (AREA)
- Traffic Control Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of ship navigation condition division, in particular to a ship navigation condition division method based on cluster analysis. The method comprises the following steps of acquiring data; removing singular data values; clustering the processed data; optimizing the clustering number; and finishing clustering analysis to obtain a clustering result. By the aid of the method, the problem that in the prior art, in the aspect of dividing ship navigation working conditions in the field of ship industry, most methods are manual experience dividing methods, errors and omission conditions are prone to occurring, ship navigation working conditions are effectively divided, and reference is provided for modeling of fuel consumption of a host.
Description
Technical Field
The invention relates to the technical field of ship navigation condition division, in particular to a ship navigation condition division method based on cluster analysis.
Background
The ship is used as a transportation tool with large transportation volume in transportation, and the operation cost of the ship is 40-60% of fuel consumption, wherein the most common ship main engine diesel engine is used as the power 'heart' of the ship, and the fuel consumption of the ship is usually more than 90% of the fuel consumption of the whole ship; taking an ocean ten thousand ton wheel as an example, the fuel consumption of a low-speed diesel engine reaches more than 20-30 tons every one day of sailing, and the operation and use cost of the ocean ten thousand ton wheel accounts for a larger proportion of the operation cost of a ship. How to reduce cost and improve efficiency becomes the key point of concern for users of various shipping enterprises; meanwhile, the fuel consumption is closely related to the emission of pollutants, and excessive fuel consumption will inevitably result in the increase of the emission of nitrogen oxides, and influence the marine atmospheric environment in the transportation sea area. Therefore, how to effectively optimize the energy saving performance and improve the environmental protection performance of ship navigation becomes a problem which is currently extremely concerned and urgently needed to be solved by shipping users.
For a marine main engine diesel engine, the modeling evaluation and optimization of the fuel consumption condition of the main engine diesel engine can be well done in the actual running process of a ship, and the method has practical significance. The navigation of a ship is divided into two distinct phases. Firstly, maneuvering sailing operating mode, also known as transition operating mode, transition operating mode boats and ships machine oar cooperation point is in the change, and such operating mode mainly includes: mooring working conditions, starting acceleration working conditions, steering working conditions, backing working conditions and the like. And secondly, under the normal (constant speed) sailing working condition, the commercial ship sails for more than 95% of the time and is under the normal sailing working condition, the sailing conditions (ship resistance and host working condition) of the ship are relatively stable, and the engine-propeller matching point of the ship does not change greatly. The main engine fuel consumption modeling is mainly characterized in that the main engine fuel consumption modeling is carried out under the normal navigation working condition, the reason influencing the main engine fuel consumption needs to be considered, and besides the influence of the main engine self factors (a fuel supply system, a scavenging and air exchange system, a combustion chamber, a timing control system and the like), the influencing factors mainly comprise:
(1) vessel draft (load);
(2) ship fouling;
(3) meteorological conditions (storm conditions);
(4) dragging a ship;
(5) narrow channel or shallow water navigation, etc.
Because the minimum fuel consumption rate of the host is different under different working conditions, different types of working conditions encountered by the host in the navigation process need to be classified before the host fuel consumption optimization sub-working condition model is established, and the next analysis is carried out on the basis of completing reasonable division of the navigation working conditions, so that a more accurate model and a more effective optimization result can be obtained. However, in the prior art, in the aspect of dividing ship navigation conditions in the field of ship industry, most methods are manual experience division methods, and error and omission conditions are easily generated.
Disclosure of Invention
In view of the above analysis, the present invention aims to provide a ship navigation condition partitioning method based on cluster analysis, so as to solve the problems that most of the prior art is divided into conditions through manual experience, and errors and omissions are easily generated.
The purpose of the invention is mainly realized by the following technical scheme:
a ship navigation condition division method based on cluster analysis is characterized by comprising the following steps:
acquiring factor data influencing fuel consumption of a host diesel engine;
the collected data are sorted;
carrying out clustering analysis on the well-regulated data vectors, processing the data through a k-means algorithm and obtaining a convergent clustering result;
and dividing the ship navigation working conditions through the obtained clustering result. The invention has the following beneficial effects: the method can truly reflect different environmental conditions of the host when the ship sails, accords with the actual process of ship operation, and can be used for measuring and calculating the actual minimum fuel consumption rate of the host and the corresponding host rotating speed and providing the basis for modeling and optimizing the fuel consumption of the host; meanwhile, the method can provide a basis for arranging and analyzing the annual operation data of the ship route in seasonal and separate navigation sections.
On the basis of the scheme, the invention is further improved as follows:
further, the factor data influencing the fuel consumption of the main engine diesel engine comprises average draft and relative wind speed data.
The beneficial effect of adopting the further scheme is that: in the process of normal navigation of the ship, through multiple data acquisition and analysis, the most important factors influencing the working condition are the average draft and the data volume of relative wind speed. The two data are collected, so that the oil consumption analysis and the working condition division can be realized, the data result is accurate and reliable, and the collection of unnecessary data is reduced.
Further, the average draft is calculated by calculating the left midship draft dL and the right midship draft dR of the ship according to a formula to obtain the average draft D of the ship, wherein D (D: (D)i)＝(dL(i)+dR(i) 2,/i represents the ith sample.
The beneficial effect of adopting the further scheme is that: by taking the average values of the left midship draft dL and the right midship draft dR of the ship, the obtained average draft of the ship enables the data for analyzing the oil consumption of the ship to be more accurate.
Further, the method for calculating the data amount of the relative wind speed is to calculate the wind direction dWWind speed sAPreprocessing the data by formula sR(i)＝-sA(i)*cos(dW(i) A relative wind speed s with respect to the bow direction is calculatedRData, where i denotes the ith sample, sRThe positive and negative numbers can represent the direct and inverse relationship between the wind direction and the bow, the positive numbers represent the direct wind, and the negative numbers represent the inverse wind.
The beneficial effect of adopting the further scheme is that: by the direction of the wind dWWind speed sAThe calculated relative wind speed value is used as an influence factor for analyzing the working condition of the ship, so that the analysis is more accurate. The positive and negative are used for replacing the forward and backward wind, so that the obtained wind direction of the relative wind speed can be judged more visually.
Further, the data sorting comprises the step of eliminating singular values by adopting an eliminating method for setting a threshold range and cleaning the collected data.
The beneficial effect of adopting the further scheme is that: the collected data are subjected to singular value removal by a threshold setting method, so that analysis is more accurate, and the influence of abnormal values on data analysis is eliminated.
Further, the arrangement of the data also comprises the step of forming the cleaned data into a data vector form.
The beneficial effect of adopting the further scheme is that: all the collected data are combined into a vector form, and the data at the same time can be processed at the same time to prepare for processing the data by a k-means algorithm.
Further, the k-means algorithm includes determining k-m × n clusters, where m denotes the number of draught states and n denotes the number of states divided with respect to the wind speed, i.e., k operating conditions.
The beneficial effect of adopting the further scheme is that: and determining the number of the working conditions needing to be analyzed, setting the number as k, wherein the number of the working conditions is also the number of the clusters in the k-means algorithm, namely the obtained analysis result is also the data of the k working conditions.
Further, the k-means algorithm further comprises repeating the following process until convergence:
for each sample i, calculate the class to which it should belong
For each class k, the centroid of the class is recalculated:
wherein the training sample is { x(1)，...，x(m)}，c(i)Representing the class of sample i that is closest to the k classes, c(i)Is 1 to k; centroid mukRepresenting the result of the training update of the sample centroids belonging to the same class.
The beneficial effect of adopting the further scheme is that: the K-Means clustering algorithm is fast and simple, high in efficiency and scalability on a large data set, time complexity is close to linearity, and the method is suitable for mining large-scale data sets.
The division result of the sailing working conditions is obtained after the sailing segment data is analyzed by using the K-Means clustering algorithm, the range interval of each sailing working condition based on the draught and the relative wind speed can be determined by respectively solving the maximum value and the minimum value of the draught and the relative wind speed of each working condition (namely the cluster of the clustering result), and a guidance basis is provided for the analysis of a later host fuel consumption working condition model.
Further, if the convergence result of the clustering number by using the working condition does not conform to the actual working condition, changing the value of k set originally, and reusing the k-means algorithm until convergence.
The beneficial effect of adopting the further scheme is that: and analyzing the range obtained by analysis and the initial preset working condition to judge whether the expected result can be achieved, and if the range does not accord with the working condition, changing the number k value of the modified working condition. And repeating the k-means algorithm to obtain a new working condition.
Further, the convergence result is divided into k types of conditions, and the interval result is quantized.
The beneficial effect of adopting the further scheme is that: and dividing the analysis result into k working conditions and dividing corresponding intervals. Thereby providing basis for establishing a host model.
In the invention, the technical schemes can be combined with each other to realize more preferable combination schemes. Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the invention, wherein like reference numerals are used to designate like parts throughout.
FIG. 1 is a schematic diagram of a scheme for dividing ship navigation conditions by k-means cluster analysis;
FIG. 2 is a draft profile;
FIG. 3 is a relative wind velocity profile;
FIG. 4 shows preliminary clustering results;
FIG. 5 Re-clustering results
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
Example one
The invention discloses a ship navigation condition division method based on cluster analysis. The method specifically comprises the following steps:
step one, factor data influencing fuel consumption of a main engine diesel engine are obtained.
Specifically, influencing factor data is collected.
Under the condition that the main engine diesel engine and the accessory equipment, the ship body and the propeller thereof are in good states, the main influence factors of the fuel consumption of the main engine diesel engine are as follows: draft (load), bottom of the vessel, weather conditions (storm conditions), vessel drag, narrow channel or shallow water sailing, etc.
And step two, the collected data are sorted.
Specifically, the collected data is first cleaned.
Because singular values caused by sensor abnormity, ship manipulation and the like exist in the acquired data, the data are not generated by the difference of working conditions, and the data have no relevance significance to the calculation of the navigation working condition clustering process. The elimination is usually performed by a method of setting a threshold range.
Specifically, the cleaned data is then combined into a sample data set.
And thirdly, performing cluster analysis on the well-regulated data vectors, processing the data through a k-means algorithm and obtaining a convergence result.
Specifically, a navigation condition clustering number k is set, and k-means clustering algorithm calculation is carried out.
The working condition of ship navigation is not classified originally, so that the draft and relative wind speed data samples of a certain ship in a certain navigation section have no class label y and only have the feature x. Clustering belongs to unsupervised learning, and aims to find potential classes y of each sample x and put samples x of the same class y together. The class into which the clustering is required to be divided is unknown. Therefore, clustering is a process of classifying data into different classes or clusters, so that objects in the same cluster have great similarity, and objects in different clusters have great dissimilarity. The classification of the ship navigation working conditions is realized by a clustering method, and the working conditions with larger similarity can be classified into one type.
K-meThe ans algorithm is a distance-based clustering algorithm, and takes distance as an evaluation index of similarity, i.e., it is considered that the closer the distance between two objects is, the greater the similarity is. The algorithm considers clusters to be composed of closely spaced objects, and therefore targets the resulting compact and independent clusters as final targets. In the clustering problem, the training sample is { x }(1)，...，x(m)Each x(i)∈RnAnd no label y.
The K-means algorithm is to cluster the samples into K clusters (cluster). The number k of the clustered categories needs to be determined by combining understanding and analysis of services. The K-means algorithm is specifically as follows:
1. randomly selecting k cluster centroids (cluster centroids) as mu1，μ2，...，μk∈Rn。
2. The following process is repeated until convergence:
for each sample i, calculate the class to which it should belong
For each class k, the centroid of the class is recalculated.
Wherein the training sample is { x(1)，...，x(m)}，c(i)Representing the class of sample i that is closest to the k classes, c(i)Is 1 to k; centroid mukRepresenting a guess of the center point of the samples belonging to the same class. The explanation by the ship navigation condition model is that all vectors of draught and relative wind speed are gathered into k working condition types, the vectors of the draught and the relative wind speed are randomly selected as centroids of the k working conditions in the first step, the distance from each vector to each of the k centroids is calculated in the second step, and the working condition with the closest distance is selected as c(i)Thus each vector has a positionBelonging to working conditions; third step of recalculating its centroid μ for each conditionk(average all vector coordinates inside); the fourth step iterates the second and third steps until the centroid is unchanged or changes very little.
And step four, dividing the ship navigation working condition through the obtained clustering result.
According to the clustering result, real ship data in each navigation condition cluster can be obtained, and a navigation condition classification range interval can be obtained through statistical calculation.
Specifically, the divided working condition is a range interval of two parameters of ship draught and relative wind speed.
Example two
The embodiment describes the ship navigation condition division method in the first embodiment in detail.
Step 1, factor data influencing fuel consumption of a diesel engine of a host is obtained.
Preferably, the data volume including average draft and relative wind speed.
Specifically, when a large ship sails in the ocean and is in a stable state, the conditions of changes of a ship bottom cleaning state, ship dragging, narrow channel navigation or shallow water navigation and the like can be ignored. However, the draught of the ship affects the navigation resistance of the ship, changes the process, loads of a main engine and the like.
When the ship normally runs on the sailing line, the slip rate of the propeller can be changed in the external stormy weather, and certain influence is caused on the normal sailing of the ship. Parameters of ship draught and wind conditions (wind direction and wind speed, relative wind speed can be obtained from the two quantities) can obtain corresponding data in operation through a sensor loaded on a ship, and a premise is provided for data analysis of sailing conditions.
Preferably, the average draft calculation method of the ship is to take the left midship draft d of the ship as the average draft calculation methodLThe water d in the middle and the rightRAnd calculating the average draught D of the ship by a formula, wherein i represents the ith sample.
D(i)＝(dL(i)+dR(i))/2
Preferably, the relative wind speed calculation of the vesselThe calculation method is to calculate the wind direction dWWind speed sAPreprocessing the data, and calculating to obtain the relative wind speed s relative to the bow direction through a formulaRAnd data, wherein positive and negative can represent the direct and inverse relationship of the wind direction to the bow, positive represents the direct wind, negative represents the upwind, and i represents the ith sample.
sR(i)＝-sA(i)*cos(dW(i))
And 2, sorting the acquired data.
The collected data is first cleaned.
Because singular values caused by sensor abnormity, ship manipulation and the like exist in the acquired data, the data are not generated by the difference of working conditions, and the data have no relevance significance to the calculation of the navigation working condition clustering process. The elimination is usually performed by a method of setting a threshold range.
Preferably, for the average draft D of the vessel, data outside the (0, vessel type depth) range are rejected, while for the relative wind speed sRData outside the (-50, 50) range are culled.
Specifically, the cleaned data is then combined into a sample data set.
Preferably, the cleaned vessel mean draft and relative wind speed data vector [ D (i), s [ ]R(i)]And forming a sample data set.
And 3, carrying out cluster analysis on the well-regulated data vectors.
And 301, assuming a navigation condition clustering number k, and calculating a k-means clustering algorithm.
Particularly, when the system is applied to a large-sized oil tanker on a real ship, most oil tankers carry crude oil to run in a single-stroke mode, the sailing state is guaranteed for the most of the forward stroke of the oil tankers under the ballast water pressure load, and the return stroke of the oil tankers is in a full-load sailing state. As shown in fig. 2, the draft state can be divided into two states of ballasting and fully loaded. The wind conditions encountered during sailing are shown in fig. 3, without obvious features, the relative wind speed conditions are classified into three categories according to the type of seal.
Therefore, it is preferable to assume that the sailing condition clustering number k is m × n is 2 × 3 × 6, i.e., m indicates the number of types of the draught state, i.e., two types of full load and ballast, and n indicates the number of types of the relative wind speed, i.e., three types of downwind, upwind, and upwind. And (5) carrying out calculation by using a k-means clustering algorithm.
Specifically, the K-means algorithm is a distance-based clustering algorithm, and the distance is used as an evaluation index of similarity, that is, the closer the distance between two objects is, the greater the similarity of the two objects is. The algorithm considers clusters to be composed of closely spaced objects, and therefore targets the resulting compact and independent clusters as final targets. In the clustering problem, the training sample is { x }(1)，...，x(m)Each x(i)∈RnAnd no label y.
The K-means algorithm is to cluster the samples into K clusters (cluster). Where the number of categories of the cluster is 6. The K-means algorithm is specifically as follows:
1. randomly selecting 6 cluster centroids (cluster centroids) as mu1，μ2，...，μ6∈Rn。
2. The following process is repeated until convergence:
for each sample i, calculate the class to which it should belong
For each class k, the centroid of the class is recalculated.
Wherein the training sample is { x(1)，...，x(m)}，c(i)Representing the class of sample i that is closest to the 6 classes, c(i)Is 1 to 6; centroid mu6Representing a guess of the center point of the samples belonging to the same class. The explanation by the ship navigation condition model is that all draught and relative wind speed vectors are gathered into 6 working condition types, and 6 draught and relative wind speed vectors are randomly selected as 6 working condition types in the first stepThe centroids of the working conditions, the second step of calculating the distance from each vector to each of the 6 centroids, and then selecting the working condition with the closest distance as c(i)Thus, each vector has the working condition; third step of recalculating its centroid μ for each condition6(average all vector coordinates inside); the fourth step iterates the second and third steps until the centroid is unchanged or changes very little.
Specifically, in the specific application of the k-means clustering algorithm, an Euclidean distance (Euclidean distance) calculation method is adopted to describe the similarity between data individuals, and other distance calculation methods (such as manhattan distance, included angle cosine distance, correlation distance, and the like) can also be utilized.
The step can obtain the clustering analysis result of k-means when the clustering number k of all the navigation condition data samples is 6.
The analysis results obtained are shown in fig. 4, where the triangles in fig. 4 are 6 cluster centers. The method comprises the following specific steps:
| clustering center | Relative wind velocity (m/s) | Draught (m) |
| 1 | -6.91 | 9.43 |
| 2 | -5.58 | 20.06 |
| 3 | 7.79 | 9.76 |
| 4 | 3.50 | 20.03 |
| 5 | 0.67 | 9.50 |
| 6 | -15.44 | 19.44 |
Step 302, optimizing the clustering number k.
The analysis results can obtain that the samples in the 6 clustering centers do not distinguish the difference between ballast and full load, and cannot meet the practical working condition application requirements. Because the ship has higher stability under the full-load state, the influence on wind is smaller than that when the ship is ballasted.
Preferably, in order to make the classification result more reliable, the number of clusters k is increased by 1, and a new k is set to 7.
And step 303, clustering again to determine the ship navigation condition division.
And substituting the optimized clustering number k-7 into a k-means clustering algorithm for calculation to obtain a clustering analysis result, which is shown in fig. 5. The specific data are as follows:
| clustering center | Relative wind velocity (m/s) | Draught (m) |
| 1 | 3.56 | 20.025 |
| 2 | -12.35 | 9.59 |
| 3 | -14.86 | 20.09 |
| 4 | 8.03 | 9.75 |
| 5 | -5.37 | 20.06 |
| 6 | 1.38 | 9.54 |
| 7 | -4.88 | 9.38 |
According to the clustering result, the characteristics of each working condition can be clearly distinguished. The method is characterized in that the range interval of two parameters of the draught and the relative wind speed of the ship classified according to each sailing condition is counted, and a basis can be provided for establishing a sub-condition model of the fuel consumption of the host. The results are as follows:
therefore, the method divides the navigation conditions of the supertanker in the stable navigation state into the following results: (1) ballasting, forward blowing; (2) ballasting and light wind; (3) ballasting, reverse blowing; (4) ballasting and reverse strong wind; (5) full load and downwind; (6) full load, reverse, and wind; (7) full load, reverse strong wind, seven types; and results for the quantization intervals are obtained for the range of each condition type.
In summary, the embodiment of the invention provides a ship navigation condition division method based on cluster analysis, which can be used for modeling the working condition by analyzing, processing and optimizing the actual ship navigation data base line, thereby improving the fuel efficiency of a host, reducing the power cost and overcoming the defect that the manual division of the working condition is inaccurate in the prior art.
Those skilled in the art will appreciate that all or part of the flow of the method implementing the above embodiments may be implemented by a computer program, which is stored in a computer readable storage medium, to instruct related hardware. The computer readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (7)
1. A ship navigation condition division method based on cluster analysis is characterized by comprising the following steps:
acquiring factor data influencing fuel consumption of a diesel engine of a host, wherein the factor data comprises data of average draft and relative wind speed;
the collected data are sorted;
performing cluster analysis on the well-organized data vectors, processing the data through a k-means algorithm, and obtaining a convergent clustering result, wherein the k-means algorithm comprises determining k-m-x-n clusters, wherein m represents the number of draught states, and n represents the number of states divided by relative wind speed, namely k working conditions;
and dividing the ship navigation working condition into two ranges of parameters of ship draught and relative wind speed through the obtained clustering result, dividing the convergence result into k types of working conditions, and quantizing the range result.
2. The vessel voyage condition division method based on the cluster analysis according to claim 1, wherein the average draft is calculated by calculating the left midship draft dL and the right midship draft dR of the vessel according to a formula to obtain the average draft D of the vessel, wherein D (i) ═ D (D)L(i)+dR(i) 2,/i represents the ith sample.
3. The method for partitioning ship voyage working condition based on cluster analysis according to claim 1, wherein the method for calculating the data quantity of the relative wind speed is to divide the wind direction dWWind speed sAPreprocessing the data by formula sR(i)＝-sA(i)*cos(dW(i) A relative wind speed s with respect to the bow direction is calculatedRData, where i denotes the ith sample, sRThe positive and negative numbers represent the direct and inverse relationship between the wind direction and the bow, the positive numbers represent the direct wind, and the negative numbers represent the inverse wind.
4. The method for dividing the sailing condition of a ship based on cluster analysis according to claim 1, wherein the data sorting comprises eliminating singular values by adopting an elimination method for setting a threshold range to clean the collected data.
5. The method for partitioning ship voyage condition based on cluster analysis according to claim 4, wherein the data sorting further comprises the step of combining the washed data into a data vector form.
6. The method for partitioning ship voyage conditions based on cluster analysis according to claim 1, wherein the k-means algorithm further comprises repeating the following process until convergence:
for each sample i, calculate the class to which it should belong
For each class k, the centroid of the class is recalculated:
wherein the training sample is { x(1)，...，x(m)}，c(i)Representing the class of sample i that is closest to the k classes, c(i)Is 1 to k; centroid mukRepresenting the result of the training update of the sample centroids belonging to the same class.
7. The method for dividing the sailing condition of a ship based on cluster analysis according to claim 6, wherein if the convergence result of the cluster number by using the working condition does not conform to the actual working condition, the value of k set originally is changed, and the k-means algorithm is reused until convergence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711214624.3A CN108052707B (en) | 2017-11-28 | 2017-11-28 | Ship navigation condition division method based on cluster analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711214624.3A CN108052707B (en) | 2017-11-28 | 2017-11-28 | Ship navigation condition division method based on cluster analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108052707A CN108052707A (en) | 2018-05-18 |
| CN108052707B true CN108052707B (en) | 2021-09-14 |
Family
ID=62120788
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711214624.3A Active CN108052707B (en) | 2017-11-28 | 2017-11-28 | Ship navigation condition division method based on cluster analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108052707B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109711682B (en) * | 2018-12-13 | 2021-09-14 | 中国船舶工业系统工程研究院 | Data-driven ship overall state evaluation method |
| CN110147820B (en) * | 2019-04-11 | 2021-06-15 | 北京远航通信息技术有限公司 | Method, device, equipment and storage medium for recommending extra oil quantity of flight |
| 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 |
| CN113428318B (en) * | 2021-08-04 | 2022-04-15 | 南京盛航海运股份有限公司 | Ship power equipment monitoring method, device, equipment and storage medium |
| CN115169434B (en) * | 2022-06-14 | 2023-09-19 | 上海船舶运输科学研究所有限公司 | Host working condition characteristic value extraction method and system based on K-means clustering algorithm |
| CN115186005A (en) * | 2022-06-16 | 2022-10-14 | 上海船舶运输科学研究所有限公司 | Working condition division method and system for ship main engine |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103921743A (en) * | 2014-05-08 | 2014-07-16 | 长春工业大学 | Automobile running working condition judgment system and judgment method thereof |
| CN105868787A (en) * | 2016-03-31 | 2016-08-17 | 江苏大学 | Electric car driving range evaluation method based on working condition identification and fuzzy energy consumption |
| CN106709662A (en) * | 2016-12-30 | 2017-05-24 | 山东鲁能软件技术有限公司 | Electrical equipment operation condition classification method |
| CN107067722A (en) * | 2017-04-24 | 2017-08-18 | 中国汽车技术研究中心 | Novel vehicle running condition construction method |
| WO2017149526A2 (en) * | 2016-03-04 | 2017-09-08 | May Patents Ltd. | A method and apparatus for cooperative usage of multiple distance meters |
-
2017
- 2017-11-28 CN CN201711214624.3A patent/CN108052707B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103921743A (en) * | 2014-05-08 | 2014-07-16 | 长春工业大学 | Automobile running working condition judgment system and judgment method thereof |
| WO2017149526A2 (en) * | 2016-03-04 | 2017-09-08 | May Patents Ltd. | A method and apparatus for cooperative usage of multiple distance meters |
| CN105868787A (en) * | 2016-03-31 | 2016-08-17 | 江苏大学 | Electric car driving range evaluation method based on working condition identification and fuzzy energy consumption |
| CN106709662A (en) * | 2016-12-30 | 2017-05-24 | 山东鲁能软件技术有限公司 | Electrical equipment operation condition classification method |
| CN107067722A (en) * | 2017-04-24 | 2017-08-18 | 中国汽车技术研究中心 | Novel vehicle running condition construction method |
Non-Patent Citations (1)
| Title |
|---|
| 一种应用数据挖掘技术评估柴油机性的方法能;孙峰 等;《大连海事大学学报》;20170831;第43卷(第3期);正文第83-88页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108052707A (en) | 2018-05-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108052707B (en) | Ship navigation condition division method based on cluster analysis | |
| CN112906830B (en) | Automatic generation method of ship optimal route based on AIS big data | |
| CN109740460B (en) | Optical remote sensing image ship detection method based on depth residual error dense network | |
| CN113537386B (en) | Ship typical motion trail self-adaptive mining method based on improved K-Medoids clustering | |
| CN107292451B (en) | Ship speed optimization method and equipment | |
| CN113450596A (en) | Space-time DP method based on ship track feature point extraction | |
| CN111552299B (en) | Intelligent optimization management system and optimization method for wind wing navigation-aiding ship energy efficiency | |
| CN111582380B (en) | Ship track density clustering method and device based on space-time characteristics | |
| CN110598693A (en) | Ship plate identification method based on fast-RCNN | |
| CN111291149A (en) | Ship AIS space-time trajectory segmentation and mode extraction method based on course deviation | |
| CN110378239A (en) | A kind of real-time traffic marker detection method based on deep learning | |
| CN113033303B (en) | SAR image rotation ship detection implementation method based on RCIoU loss | |
| CN111914839A (en) | Synchronous end-to-end license plate positioning and identifying method based on YOLOv3 | |
| CN110991516A (en) | Side-scan sonar image target classification method based on style migration | |
| CN113157732A (en) | Underground scraper fault diagnosis method based on PSO-BP neural network | |
| CN109102538B (en) | Method for extracting land characteristic points and land characteristic lines by using contour line data and weather map analysis method | |
| CN113836118A (en) | Ship static data supplementing method and device, electronic equipment and readable storage medium | |
| CN114254438A (en) | Dynamic rotating speed fine adjustment method for optimizing ship speed | |
| CN117036993A (en) | Ship water gauge remote measurement method based on unmanned aerial vehicle | |
| Zou et al. | Maritime target detection of intelligent ship based on faster R-CNN | |
| CN114048798A (en) | Automobile driving condition construction method based on improved noise reduction self-encoder | |
| Fu et al. | Ship detection based on Improved YOLO Algorithm | |
| CN116861339A (en) | Method for improving estimation accuracy of carbon emission of ship by identifying ship working conditions based on machine learning | |
| CN115731211A (en) | Method for detecting faults of contact net fastener by adopting four-stage cascade architecture | |
| CN110287763A (en) | A kind of candidate frame ratio optimization method towards ship seakeeping application |
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 |