CN115169434B - Host working condition characteristic value extraction method and system based on K-means clustering algorithm - Google Patents
Host working condition characteristic value extraction method and system based on K-means clustering algorithm Download PDFInfo
- Publication number
- CN115169434B CN115169434B CN202210665119.5A CN202210665119A CN115169434B CN 115169434 B CN115169434 B CN 115169434B CN 202210665119 A CN202210665119 A CN 202210665119A CN 115169434 B CN115169434 B CN 115169434B
- Authority
- CN
- China
- Prior art keywords
- host
- average value
- working condition
- characteristic
- clustering algorithm
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Feedback Control In General (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention provides a host operating condition characteristic value extraction method and system based on a K-means clustering algorithm, which comprises the steps of firstly collecting ship data and preprocessing, selecting a plurality of characteristic parameters from the preprocessed ship data, calculating to obtain the average value of each characteristic parameter in unit time, then adopting an F test method to extract the average value of the characteristic parameter with the maximum correlation with the host power average value in the characteristic parameters as the average value of the related characteristic parameters, adopting a K-means clustering algorithm to conduct first-stage operating condition division on the host power average value and the average value of the related characteristic parameters to obtain a plurality of host operating conditions, adopting the K-means clustering algorithm to conduct second-stage operating condition division on the host operating condition with the lowest contour coefficient by combining the sea water temperature average value and the host scavenging box temperature average value, finally respectively extracting the characteristic values of the divided host operating conditions, and improving the accuracy of the host operating condition division.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a host operating condition characteristic value extraction method and system based on a K-means clustering algorithm.
Background
The daily operating costs of the ship are very high and the effective operating time determines the level of profits of the shipper. The ship main engine is a power device of a ship core, and faults can be inevitably generated in the using process, the occurrence of the faults can seriously influence the normal operation of the ship, not only the normal operation of equipment, but also accidents can be generated when the faults are serious, and even the personal safety is endangered.
The performance of the ship main engine is an important index of the ship navigation state. In the field of intelligent ship big data mining operation, the present stage is lack of effective quantitative data classification when analyzing the host performance in ship navigation, the host is usually only modeled and optimized, but under different working conditions, the performance standard of the host is different, so that the division of the working conditions of the host has important significance for the research of the performance of the host. The working condition dividing method existing at the present stage is single, is mostly based on single-layer single-stage division, and has certain limitation in practical application, particularly in mass real ship operation data analysis.
The division of the working conditions of the host is a main basis for realizing the efficiency, the fault prediction and the operation analysis of the host. According to ship equipment information and navigation information acquired by a real ship, the working condition of a host is reasonably divided according to a physical prototype of the operation of the host, and a plurality of equipment parameters including parameters of a supercharger, parameters of cooling water and other complex mechanisms are required to be considered.
In practical engineering applications, many equipment parameters are difficult to obtain and sometimes data is lost. The change of the working condition of the main engine is also affected by temperature and machine aging, for example, the working condition of the main engine is different between the period of the ship operation and the period of the ship operation for many years, so that the reasonable operation working condition of the main engine of the ship is divided according to the real ship data, a foundation is laid for determining the pollutant discharge amount of the ship, estimating the fuel consumption amount, estimating the performance evaluation of the main engine, diagnosing and predicting the fault of key equipment of the main engine, and the like, and a reference basis is provided for the management and maintenance of the ship equipment.
The operation of the working condition of the host is a coupling process, and all main devices can mutually influence to change the working condition, but the actual ship can hardly comprehensively output parameters required by the physical model of the operation of the host, so that the working condition division has certain difficulty according to the existing method.
Disclosure of Invention
In order to solve the problems that the dividing method is single in the existing host working condition dividing process, and has larger limitation in practical application and mass real ship operation data analysis, the invention provides a host working condition characteristic value extracting method based on a K-means clustering algorithm, which is used for carrying out correlation analysis based on an F test method, carrying out working condition division on host power and characteristic parameters with highest correlation by adopting the K-means clustering algorithm, calculating the contour coefficient by adopting a specific calculating method, and continuing to carry out working condition division on the host working condition with the lowest contour coefficient, thereby effectively improving the working condition identification and the availability of practical engineering, simultaneously effectively dividing the ship host working condition by adopting two-stage division on the host working condition, improving the accuracy of the host working condition division, and providing a basis for further analyzing and optimizing the working performance of ships and hosts. The invention also relates to a host working condition characteristic value extraction system based on the K-means clustering algorithm.
The technical scheme of the invention is as follows:
the host computer working condition characteristic value extraction method based on the K-means clustering algorithm is characterized by comprising the following steps of:
and a data acquisition and processing step: the method comprises the steps of collecting ship data, preprocessing the ship data, and selecting a plurality of characteristic parameters from the preprocessed ship data, wherein the characteristic parameters comprise host power, sea water temperature and host scavenging box temperature; calculating to obtain the average value of each characteristic parameter in unit time, wherein the average value comprises a main engine power average value, a sea water temperature average value and a main engine scavenging box temperature average value;
the first working condition dividing step: extracting a characteristic parameter average value with the greatest correlation with a host power average value by adopting an F test method, taking the characteristic parameter average value as a correlated characteristic parameter average value, and carrying out first-stage working condition division on the host power average value and the correlated characteristic parameter average value by adopting a K-means clustering algorithm to obtain a plurality of host working conditions;
the second working condition dividing step: calculating the corresponding contour coefficient under each host working condition, and then adopting a K-means clustering algorithm to continuously divide the host working condition with the lowest contour coefficient by combining the sea water temperature average value and the host scavenging box temperature average value to obtain a plurality of host working conditions, wherein the number of host working conditions obtained in the second-stage working condition dividing step is the same as or different from that of the host working conditions obtained in the first-stage working condition dividing step;
and a characteristic value extraction step: and respectively calculating characteristic values of the working conditions of the host machine obtained in the first working condition dividing step and the second working condition dividing step, and integrating calculation results to analyze the working performance of the host machine.
Preferably, in the second working condition dividing step, when the second stage working condition dividing is performed, an elbow method is further adopted to calculate the optimal number of host working conditions, and the optimal number of host working conditions is used as the final obtained number of host working conditions; the characteristic value extraction step calculates the characteristic value of each optimized host working condition.
Preferably, in the data acquisition processing step, the preprocessing includes deleting abnormal data in the ship data, and removing noise and data normalization.
Preferably, in the data acquisition and processing step, the characteristic parameters further include a host rotation speed, a ship navigational speed, a host cylinder exhaust temperature and a host cylinder cooling water temperature.
Preferably, in the step of extracting the characteristic value, the characteristic value of each host operating condition includes an average value and a standard deviation of host power, an average value and a standard deviation of host rotational speed, an average value and a standard deviation of ship speed, an average value and a standard deviation of sea water temperature, an average value and a standard deviation of host cylinder exhaust temperature, an average value and a standard deviation of host cylinder cooling water temperature, and an average value and a standard deviation of host scavenging box temperature, which are all calculated under each host operating condition.
A host working condition characteristic value extraction system based on a K-means clustering algorithm is characterized by comprising a data acquisition processing module, a first working condition dividing module, a second working condition dividing module and a characteristic value extraction module which are connected in sequence,
the data acquisition and processing module: the method comprises the steps of collecting ship data, preprocessing the ship data, and selecting a plurality of characteristic parameters from the preprocessed ship data, wherein the characteristic parameters comprise host power, sea water temperature and host scavenging box temperature; calculating to obtain the average value of each characteristic parameter in unit time, wherein the average value comprises a main engine power average value, a sea water temperature average value and a main engine scavenging box temperature average value;
the first working condition dividing module: extracting a characteristic parameter average value with the greatest correlation with a host power average value by adopting an F test method, taking the characteristic parameter average value as a correlated characteristic parameter average value, and carrying out first-stage working condition division on the host power average value and the correlated characteristic parameter average value by adopting a K-means clustering algorithm to obtain a plurality of host working conditions;
the second working condition dividing module: calculating the corresponding contour coefficient under each host working condition, and then adopting a K-means clustering algorithm to continuously divide the host working condition with the lowest contour coefficient by combining the sea water temperature average value and the host scavenging box temperature average value to obtain a plurality of host working conditions, wherein the number of host working conditions obtained by dividing the second stage working conditions is the same as or different from that of the host working conditions obtained by dividing the first stage working conditions;
the characteristic value extraction module: and respectively calculating characteristic values of the working conditions of the host machine obtained in the first working condition dividing module and the second working condition dividing module, and integrating calculation results to analyze the working performance of the host machine.
Preferably, in the second working condition dividing module, when the second stage working condition dividing is performed, an elbow method is further adopted to calculate the optimal number of host working conditions, and the optimal number of host working conditions is used as the final obtained number of host working conditions; the characteristic value extraction module calculates the characteristic value of each optimized host working condition.
Preferably, the preprocessing includes deleting abnormal data in the ship data, and removing noise and data normalization.
Preferably, the characteristic parameters further comprise a host rotational speed, a ship navigational speed, a host cylinder exhaust temperature and a host cylinder cooling water temperature.
Preferably, the characteristic values of each host working condition comprise an average value and a standard deviation of host power, an average value and a standard deviation of host rotating speed, an average value and a standard deviation of ship navigational speed, an average value and a standard deviation of sea water temperature, an average value and a standard deviation of host cylinder exhaust temperature, an average value and a standard deviation of host cylinder cooling water temperature, and an average value and a standard deviation of host scavenging box temperature which are all calculated under each host working condition.
The beneficial effects of the invention are as follows:
according to the host working condition characteristic value extraction method based on the K-means clustering algorithm (called K-means clustering algorithm for short), correlation analysis is carried out on the average value of each characteristic parameter in ship data and the average value of host power based on the F-test method, the K-means clustering algorithm is adopted to carry out first-stage working condition division on the host power and the characteristic parameter with the highest correlation, a plurality of host working conditions are divided, the contour coefficient of the host working conditions is calculated, and second-stage working condition division, namely division, is carried out again on the host working condition with the lowest contour coefficient, so that the working condition identification and the usability of actual engineering can be effectively improved, and the characteristic value of each working condition is calculated. The method is easy to realize by adopting a K-means clustering algorithm, an F test method and the like, and in the application of a real ship, the operation condition of the ship host can be truly and effectively divided by adopting two-stage division on the host working condition, the accuracy of the division of the host working condition is improved, and a foundation can be provided for further analyzing and optimizing the working performances of the ship and the host.
The invention also relates to a host working condition characteristic value extraction system based on a K-means clustering algorithm (K-means clustering algorithm for short), which corresponds to the host working condition characteristic value extraction method based on the K-means clustering algorithm, and can be understood as a system for realizing the host working condition characteristic value extraction method based on the K-means clustering algorithm.
Drawings
FIG. 1 is a flow chart of a host operating mode characteristic value extraction method based on a K-means clustering algorithm.
FIG. 2 is a preferred flow chart of the host operating mode characteristic value extraction method based on the K-means clustering algorithm.
Detailed Description
The present invention will be described below with reference to the accompanying drawings.
The invention relates to a host operating mode characteristic value extraction method based on a K-means clustering algorithm (K-means clustering algorithm for short), and a flow chart of the method is shown in figure 1, and the method sequentially comprises the following steps:
a data acquisition processing step, or further called a data acquisition and ship data preprocessing step, wherein ship data is firstly acquired, the ship data is preprocessed, abnormal data (some data are missing in the data acquisition process, singular data are generated, the data are removed and processed), noise and data standardization are removed, and then a plurality of characteristic parameters are selected from the preprocessed ship data, wherein the characteristic parameters preferably comprise host power, sea water temperature, host scavenging box temperature, host rotating speed, ship navigational speed, host cylinder exhaust temperature, host cylinder cooling water temperature and the like; the average value of each characteristic parameter in unit time (for example, ten minutes) is calculated and obtained, and the average value comprises a main engine power average value, a sea water temperature average value, a main engine scavenging box temperature average value, a main engine rotating speed average value, a ship navigational speed average value, a main engine cylinder exhaust temperature average value, a main engine cylinder cooling water temperature average value and the like.
And then dividing the working conditions of the ship host at two stages based on a K-means clustering algorithm. A first working condition dividing step, namely first stage working condition dividing, wherein a characteristic parameter average value with the largest correlation with a host power average value is extracted by adopting an F test method (abbreviated as an F-test method) to serve as a relevant characteristic parameter average value, and a K-means clustering algorithm is adopted to carry out first stage working condition dividing on the host power average value and the relevant characteristic parameter average value to obtain a plurality of host working conditions;
specifically, as shown in the preferred flowchart of fig. 2, an F-test method is first used to extract a feature parameter average value with the greatest correlation with the host power average value from a plurality of feature parameters, and the F-test method is calculated according to the following formula:
in the above formula, F is a calculation result score value, x 1 For any characteristic parameter average value other than the host power average value, x 2 For the average value of the host power, n 1 Representing the mean value x of characteristic parameters 1 Number of x i1 Represents x 1 Ith value, n 2 Representing the average value x of the host power 2 Number of x i2 Represents x 2 The ith value, score value F, is calculated from any characteristic parameter mean value x other than the host power mean value 1 Is the calculated result value S of (2) 1 Power x with host 2 Is the calculated result value S of (2) 2 And the preparation method is compared with the prior art.
Substituting the host power average value, the sea water temperature average value, the host scavenging box temperature average value, the host rotating speed average value, the ship navigational speed average value, the host cylinder exhaust temperature average value and the host cylinder cooling water temperature average value into the above formula in sequence to calculate to obtain a characteristic parameter A with the highest score value F as a related characteristic parameter average value, and carrying out first-stage working condition division on the host power average value and the related characteristic parameter average value by adopting a K-means clustering algorithm to divide a plurality of host working conditions; the host working conditions are preferably four, that is, the host working condition data obtained by dividing the first stage working conditions are determined to be four, for example, the characteristic parameter a is divided into a high value and a low value, and the average value of the host power is also divided into a high value and a low value, so that four combinations exist, that is, four host working conditions.
And a second working condition dividing step, namely second stage working condition dividing, of calculating the corresponding contour coefficient under each host working condition, and then adopting a K-means clustering algorithm to continuously divide the host working condition with the lowest contour coefficient by combining the average value of the sea water temperature and the average value of the temperature of the host scavenging box to obtain a plurality of host working conditions, wherein the number of the host working conditions obtained in the second stage working condition dividing step is the same as or different from that of the host working conditions obtained in the first stage working condition dividing step.
Specifically, according to the four host operating conditions divided in the first operating condition dividing step, calculating a profile coefficient SC of a corresponding data cluster under each host operating condition, wherein the profile coefficient SC is calculated according to the following formula:
in the above formula, the data point I belongs to the data cluster I (namely corresponding to the working condition), a (I) represents the average value of the distances from the data point I to other data points, C I Indicating that data cluster I has C I Data point, j represents the jth data point in data cluster I, and d (I, j) represents the distance between I and j
In the above formula, b (i) represents the minimum value of the average value of the distances from the data point i to other data points, C J Indicating that data cluster J has C J Data point J represents the jth data point in data cluster J, and d (i, J) represents the distance between i and J.
In the above equation, S (i) represents the calculation coefficient of the data point i.
In the above formula, SC represents the contour coefficient of the data cluster I.
After calculating the corresponding contour coefficient SC under each host working condition, adopting a K-means clustering algorithm, taking a sea water temperature average value and a host scavenging box temperature average value as input parameters of the K-means clustering algorithm, continuing to divide the host working condition with the lowest contour coefficient into second-stage working conditions to obtain K host working conditions, adopting an elbow method to calculate the optimal host working condition number n, and taking the optimal host working condition number n as the final host working condition number (K=n), wherein K epsilon 1,2,3,4, n and m are sequentially brought into an elbow method calculation formula to calculate the optimal host working condition number n, the host working condition number finally obtained by the second-stage working condition division is preferably four (n=4), and dividing the host working condition with the lowest contour coefficient again can improve the identification of the host working conditions and the usability of actual engineering. The calculation formula of the elbow method is as follows:
in the above, C i Representing the center of each data cluster, p is the index of C i The data points of the data clusters formed by the centers are represented by n. The most suitable working condition number selected by the elbow method can reduce blindness of secondary division working conditions, and reliability of the working conditions is improved.
And a characteristic value extraction step: and respectively calculating characteristic values of the working conditions of the host machine obtained in the first working condition dividing step and the second working condition dividing step, and integrating calculation results to analyze the working performance of the host machine. Preferably, the characteristic value comprises a series of average values and standard deviations calculated under each host operating condition: the average value and standard deviation of the power of the main engine, the average value and standard deviation of the rotating speed of the main engine, the average value and standard deviation of the navigational speed of the ship, the average value and standard deviation of the sea water temperature, the average value and standard deviation of the exhaust temperature of the main engine cylinder, the average value and standard deviation of the cooling water temperature of the main engine cylinder and the average value and standard deviation of the temperature of the scavenging box of the main engine.
The average value of the host power, the average value of the host rotation speed, the average value of the ship navigational speed, the average value of the sea water temperature, the average value of the host cylinder exhaust temperature, the average value of the host cylinder cooling water temperature and the average value of the host scavenging box temperature in the characteristic values of all the host working conditions are all the average values calculated by the average value of all the characteristic parameters in the data acquisition and processing step.
When four host working conditions are divided in the first working condition dividing step, the secondary working conditions of one host working condition with the lowest profile coefficient are divided into n finally, so that the total number of host working conditions obtained in the first working condition dividing step and the second working condition dividing step is 3+n, when n=4, namely, the total number of host working conditions obtained by two stage division is 7, the average value and standard deviation of host power under the 3+n working conditions (such as 7) are calculated, the average value and standard deviation of host rotating speed, the average value and standard deviation of ship navigational speed, the average value and standard deviation of sea water temperature, the average value and standard deviation of host cylinder exhaust temperature, the average value and standard deviation of host cylinder cooling water temperature and the average value and standard deviation of host scavenging box temperature are integrated, and the calculation results are integrated to carry out subsequent host performance analysis, so that the specific conditions of all host working conditions are analyzed. For example: the method comprises the steps that parameters such as host power, host rotation speed, ship navigational speed, sea water temperature, host air cylinder exhaust temperature, host air cylinder cooling water temperature and host scavenging box temperature are arranged under each host working condition, after the average value and standard deviation of each parameter are calculated, the maximum value and the minimum value of the average value and the standard deviation of each parameter can be obtained, the working condition of the current host operation can be analyzed according to the maximum value and the minimum value of the average value and the standard deviation of each parameter, and therefore the difference between the current host operation parameter and the historical characteristic value under the working condition is compared, and whether the current host fails or the performance is reduced is judged.
The invention also relates to a host operating mode characteristic value extraction system based on the K-means clustering algorithm, which corresponds to the host operating mode characteristic value extraction method based on the K-means clustering algorithm, and can be understood as a system for realizing the method, and the system comprises a data acquisition processing module, a first operating mode dividing module, a second operating mode dividing module and a characteristic value extraction module which are connected in sequence, in particular,
the data acquisition processing module acquires ship data, performs pretreatment on the ship data, and selects a plurality of characteristic parameters from the pretreated ship data, wherein the characteristic parameters comprise host power, sea water temperature and host scavenging box temperature; calculating to obtain the average value of each characteristic parameter in unit time, wherein the average value comprises a main engine power average value, a sea water temperature average value and a main engine scavenging box temperature average value;
the first working condition dividing module is used for extracting a characteristic parameter average value with the maximum correlation with the host power average value by adopting an F test method, taking the characteristic parameter average value as a correlated characteristic parameter average value, and carrying out first-stage working condition division on the host power average value and the correlated characteristic parameter average value by adopting a K-means clustering algorithm to obtain a plurality of host working conditions;
the second working condition dividing module is used for calculating the corresponding contour coefficient under each host working condition, then adopting a K-means clustering algorithm, and combining the average value of the sea water temperature and the average value of the temperature of the host scavenging box to continuously divide the host working condition with the lowest contour coefficient into a plurality of host working conditions, wherein the number of the host working conditions obtained by the second-stage working condition division is the same as or different from that of the host working conditions obtained by the first-stage working condition division;
and the characteristic value extraction module is used for respectively calculating characteristic values of the working conditions of the host machine obtained in the first working condition division module and the second working condition division module, and integrating calculation results to analyze the working performance of the host machine.
Preferably, in the second working condition dividing module, when the second-stage working condition dividing is performed, the elbow method is further adopted to calculate the optimal number of host working conditions, and the optimal number of host working conditions is used as the final obtained number of host working conditions; the characteristic value extraction module calculates the characteristic value of each optimized host working condition.
Preferably, the preprocessing includes deleting abnormal data in the ship data, and removing noise and data normalization.
Preferably, the characteristic parameters further include a host rotational speed, a vessel navigational speed, a host cylinder exhaust temperature, and a host cylinder cooling water temperature.
Preferably, the characteristic values of each host working condition comprise an average value and a standard deviation of host power, an average value and a standard deviation of host rotating speed, an average value and a standard deviation of ship navigational speed, an average value and a standard deviation of sea water temperature, an average value and a standard deviation of host cylinder exhaust temperature, an average value and a standard deviation of host cylinder cooling water temperature, and an average value and a standard deviation of host scavenging box temperature which are all calculated under each host working condition.
The invention provides an objective and scientific host working condition characteristic value extraction method and system based on a K-means clustering algorithm, which are used for carrying out correlation analysis based on an F-test detection method, carrying out working condition division by adopting the K-means clustering algorithm, calculating a contour coefficient by adopting a specific calculation method, continuing working condition division of a host working condition with the lowest contour coefficient, effectively improving the working condition identification and the usability of actual engineering, simultaneously effectively dividing the operation working condition of a ship host by adopting two-stage division of the host working condition, improving the accuracy of the division of the host working condition, and providing a basis for further analyzing and optimizing the working performance of the ship and the host.
It should be noted that the above-described embodiments will enable those skilled in the art to more fully understand the invention, but do not limit it in any way. Therefore, although the present invention has been described in detail with reference to the drawings and examples, it will be understood by those skilled in the art that the present invention may be modified or equivalent, and in all cases, all technical solutions and modifications which do not depart from the spirit and scope of the present invention are intended to be included in the scope of the present invention.
Claims (10)
1. The host computer working condition characteristic value extraction method based on the K-means clustering algorithm is characterized by comprising the following steps of:
and a data acquisition and processing step: the method comprises the steps of collecting ship data, preprocessing the ship data, and selecting a plurality of characteristic parameters from the preprocessed ship data, wherein the characteristic parameters comprise host power, sea water temperature and host scavenging box temperature; calculating to obtain the average value of each characteristic parameter in unit time, wherein the average value comprises a main engine power average value, a sea water temperature average value and a main engine scavenging box temperature average value;
the first working condition dividing step: extracting a characteristic parameter average value with the greatest correlation with a host power average value by adopting an F test method, taking the characteristic parameter average value as a correlated characteristic parameter average value, and carrying out first-stage working condition division on the host power average value and the correlated characteristic parameter average value by adopting a K-means clustering algorithm to obtain a plurality of host working conditions;
the second working condition dividing step: calculating the corresponding contour coefficient under each host working condition, and then adopting a K-means clustering algorithm to continuously divide the host working condition with the lowest contour coefficient by combining the sea water temperature average value and the host scavenging box temperature average value to obtain a plurality of host working conditions, wherein the number of host working conditions obtained in the second-stage working condition dividing step is the same as or different from that of the host working conditions obtained in the first-stage working condition dividing step;
and a characteristic value extraction step: and respectively calculating characteristic values of the working conditions of the host machine obtained in the first working condition dividing step and the second working condition dividing step, and integrating calculation results to analyze the working performance of the host machine.
2. The method for extracting the characteristic values of the host operating conditions based on the K-means clustering algorithm according to claim 1, wherein in the second operating condition dividing step, when the second stage operating condition dividing is performed, the optimal number of host operating conditions is calculated by an elbow method, and the optimal number of host operating conditions is used as the finally obtained number of host operating conditions; the characteristic value extraction step calculates the characteristic value of each optimized host working condition.
3. The method for extracting the characteristic value of the host operating mode based on the K-means clustering algorithm according to claim 1, wherein in the data acquisition and processing step, the preprocessing includes deleting abnormal data in ship data, removing noise and normalizing the data.
4. The method for extracting the characteristic values of the working conditions of the host based on the K-means clustering algorithm according to claim 1, wherein in the data acquisition and processing step, the characteristic parameters further comprise a host rotation speed, a ship navigation speed, a host cylinder exhaust temperature and a host cylinder cooling water temperature.
5. The method for extracting the characteristic values of the host operating conditions based on the K-means clustering algorithm according to claim 4, wherein in the characteristic value extracting step, the characteristic values of each host operating condition comprise an average value and a standard deviation of host power, an average value and a standard deviation of host rotating speed, an average value and a standard deviation of ship speed, an average value and a standard deviation of sea water temperature, an average value and a standard deviation of host cylinder exhaust temperature, an average value and a standard deviation of host cylinder cooling water temperature, and an average value and a standard deviation of host scavenging box temperature which are all calculated under each host operating condition.
6. A host working condition characteristic value extraction system based on a K-means clustering algorithm is characterized by comprising a data acquisition processing module, a first working condition dividing module, a second working condition dividing module and a characteristic value extraction module which are connected in sequence,
the data acquisition and processing module: the method comprises the steps of collecting ship data, preprocessing the ship data, and selecting a plurality of characteristic parameters from the preprocessed ship data, wherein the characteristic parameters comprise host power, sea water temperature and host scavenging box temperature; calculating to obtain the average value of each characteristic parameter in unit time, wherein the average value comprises a main engine power average value, a sea water temperature average value and a main engine scavenging box temperature average value;
the first working condition dividing module: extracting a characteristic parameter average value with the greatest correlation with a host power average value by adopting an F test method, taking the characteristic parameter average value as a correlated characteristic parameter average value, and carrying out first-stage working condition division on the host power average value and the correlated characteristic parameter average value by adopting a K-means clustering algorithm to obtain a plurality of host working conditions;
the second working condition dividing module: calculating the corresponding contour coefficient under each host working condition, and then adopting a K-means clustering algorithm to continuously divide the host working condition with the lowest contour coefficient by combining the sea water temperature average value and the host scavenging box temperature average value to obtain a plurality of host working conditions, wherein the number of host working conditions obtained by dividing the second stage working conditions is the same as or different from that of the host working conditions obtained by dividing the first stage working conditions;
the characteristic value extraction module: and respectively calculating characteristic values of the working conditions of the host machine obtained in the first working condition dividing module and the second working condition dividing module, and integrating calculation results to analyze the working performance of the host machine.
7. The host computer working condition characteristic value extraction system based on the K-means clustering algorithm according to claim 6, wherein in the second working condition dividing module, when the second-stage working condition dividing is performed, an elbow method is further adopted to calculate the optimal number of host computer working conditions, and the optimal number of host computer working conditions is used as the final obtained number of host computer working conditions; the characteristic value extraction module calculates the characteristic value of each optimized host working condition.
8. The host operating mode characteristic value extraction system based on the K-means clustering algorithm according to claim 6, wherein the preprocessing includes deleting abnormal data in ship data, and removing noise and data normalization.
9. The host operating condition characteristic value extraction system based on the K-means clustering algorithm according to claim 6, wherein the characteristic parameters further comprise host rotation speed, ship speed, host cylinder exhaust temperature and host cylinder cooling water temperature.
10. The host operating condition characteristic value extraction system based on the K-means clustering algorithm according to claim 9, wherein the characteristic values of each host operating condition comprise an average value and a standard deviation of host power, an average value and a standard deviation of host rotating speed, an average value and a standard deviation of ship navigational speed, an average value and a standard deviation of sea water temperature, an average value and a standard deviation of host cylinder exhaust temperature, an average value and a standard deviation of host cylinder cooling water temperature, and an average value and a standard deviation of host scavenging box temperature which are all calculated under each host operating condition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210665119.5A CN115169434B (en) | 2022-06-14 | 2022-06-14 | Host working condition characteristic value extraction method and system based on K-means clustering algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210665119.5A CN115169434B (en) | 2022-06-14 | 2022-06-14 | Host working condition characteristic value extraction method and system based on K-means clustering algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115169434A CN115169434A (en) | 2022-10-11 |
| CN115169434B true CN115169434B (en) | 2023-09-19 |
Family
ID=83486289
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210665119.5A Active CN115169434B (en) | 2022-06-14 | 2022-06-14 | Host working condition characteristic value extraction method and system based on K-means clustering algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115169434B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012089926A1 (en) * | 2010-12-31 | 2012-07-05 | Eniram Oy | A method for determining statistical distribution of characteristic parameters of a vessel |
| CN105930629A (en) * | 2016-07-13 | 2016-09-07 | 大唐南京环保科技有限责任公司 | On-line fault diagnosis method based on massive amounts of operating data |
| CN106991524A (en) * | 2017-03-20 | 2017-07-28 | 国网江苏省电力公司常州供电公司 | A kind of platform area line loss per unit predictor method |
| CN108052707A (en) * | 2017-11-28 | 2018-05-18 | 中国船舶工业系统工程研究院 | A kind of ship's navigation operating mode division methods based on cluster analysis |
| CN112270312A (en) * | 2020-11-26 | 2021-01-26 | 中南林业科技大学 | Fan bearing fault diagnosis method and system, computer equipment and storage medium |
| CN112307688A (en) * | 2020-10-29 | 2021-02-02 | 中国舰船研究设计中心 | System and method for monitoring, managing and optimizing energy efficiency of electric propulsion ship |
| CN113128120A (en) * | 2021-04-22 | 2021-07-16 | 三一汽车起重机械有限公司 | Method for constructing typical running condition of automobile crane |
| CN113240201A (en) * | 2021-06-08 | 2021-08-10 | 兰州大学 | Method for predicting ship host power based on GMM-DNN hybrid model |
| CN113537386A (en) * | 2021-08-01 | 2021-10-22 | 大连海事大学 | Ship typical motion track self-adaptive mining method based on improved K-Medoids clustering |
| CN114254438A (en) * | 2021-12-23 | 2022-03-29 | 震兑工业智能科技有限公司 | Dynamic rotating speed fine adjustment method for optimizing ship speed |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9589045B2 (en) * | 2014-04-08 | 2017-03-07 | International Business Machines Corporation | Distributed clustering with outlier detection |
| CN111401573B (en) * | 2018-12-17 | 2023-10-27 | 中国科学院沈阳自动化研究所 | Working condition state modeling and model correcting method |
-
2022
- 2022-06-14 CN CN202210665119.5A patent/CN115169434B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012089926A1 (en) * | 2010-12-31 | 2012-07-05 | Eniram Oy | A method for determining statistical distribution of characteristic parameters of a vessel |
| CN105930629A (en) * | 2016-07-13 | 2016-09-07 | 大唐南京环保科技有限责任公司 | On-line fault diagnosis method based on massive amounts of operating data |
| CN106991524A (en) * | 2017-03-20 | 2017-07-28 | 国网江苏省电力公司常州供电公司 | A kind of platform area line loss per unit predictor method |
| CN108052707A (en) * | 2017-11-28 | 2018-05-18 | 中国船舶工业系统工程研究院 | A kind of ship's navigation operating mode division methods based on cluster analysis |
| CN112307688A (en) * | 2020-10-29 | 2021-02-02 | 中国舰船研究设计中心 | System and method for monitoring, managing and optimizing energy efficiency of electric propulsion ship |
| CN112270312A (en) * | 2020-11-26 | 2021-01-26 | 中南林业科技大学 | Fan bearing fault diagnosis method and system, computer equipment and storage medium |
| CN113128120A (en) * | 2021-04-22 | 2021-07-16 | 三一汽车起重机械有限公司 | Method for constructing typical running condition of automobile crane |
| CN113240201A (en) * | 2021-06-08 | 2021-08-10 | 兰州大学 | Method for predicting ship host power based on GMM-DNN hybrid model |
| CN113537386A (en) * | 2021-08-01 | 2021-10-22 | 大连海事大学 | Ship typical motion track self-adaptive mining method based on improved K-Medoids clustering |
| CN114254438A (en) * | 2021-12-23 | 2022-03-29 | 震兑工业智能科技有限公司 | Dynamic rotating speed fine adjustment method for optimizing ship speed |
Non-Patent Citations (4)
| Title |
|---|
| A differential privacy protecting K-means clustering algorithm based on contour coefficients;Yaling Zhang, Na Liu, Shangping Wang;《 PLoS ONE》;1-15 * |
| Data-driven prognostics using a combination of constrained K-means clustering, fuzzy modeling and LOF-based score;Alberto Diez-Olivan等;《Neurocomputing》;97-107 * |
| 基于K-means聚类算法的数据分析模型应用研究;沈泓;刘顺;;软件导刊(03);103-107 * |
| 聚类算法在船舶能效数据挖掘中的应用;高梓博;杜太利;张勇;黄连忠;;武汉理工大学学报(交通科学与工程版)(02);286-290 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115169434A (en) | 2022-10-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107784325B (en) | Spiral fault diagnosis method based on data-driven incremental fusion | |
| CN114282579A (en) | Aviation bearing fault diagnosis method based on variational modal decomposition and residual error network | |
| CN110008898B (en) | Industrial equipment data edge processing method based on symbol and convolutional neural network | |
| CN114386537B (en) | Lithium battery fault diagnosis method and device based on Catboost and electronic equipment | |
| CN115187832A (en) | Energy system fault diagnosis method based on deep learning and gram angular field image | |
| CN115638875B (en) | Power plant equipment fault diagnosis method and system based on map analysis | |
| CN111190349A (en) | Method, system and medium for monitoring state and diagnosing fault of ship engine room equipment | |
| CN112207631B (en) | Method for generating tool detection model, method, system, device and medium for detecting tool detection model | |
| CN113918642A (en) | Data filtering, monitoring and early warning method based on power Internet of things equipment | |
| CN114878172B (en) | Ship host cylinder temperature signal feature extraction method and system | |
| CN110220885B (en) | Comprehensive judgment method for wear state of mechanical equipment | |
| CN116796271A (en) | Resident energy abnormality identification method | |
| CN115375026A (en) | Method for predicting service life of aircraft engine in multiple fault modes | |
| Mi et al. | A nonparametric cumulative sum-based fault detection method for rolling bearings using high-level extended isolated forest | |
| CN114974306A (en) | Transformer abnormal voiceprint detection and identification method and device based on deep learning | |
| CN115169434B (en) | Host working condition characteristic value extraction method and system based on K-means clustering algorithm | |
| CN117272753A (en) | Power transmission line icing galloping early warning method based on process parameters | |
| CN114994451B (en) | Ship electrical equipment fault detection method and system | |
| CN114997332A (en) | Fault diagnosis method for wavelet packet analysis and sparrow algorithm optimization correlation vector machine | |
| CN114969638A (en) | Bridge performance abnormity early warning method based on modal equivalent standardization | |
| CN112883639A (en) | GIS equipment service life prediction device and method based on machine learning | |
| Jia et al. | A Deep Learning Combination Model to Predict TBM Disc-cutter Wear Status | |
| CN117420345B (en) | Power grid operation abnormal state monitoring system based on data driving | |
| CN115099319A (en) | Method and system for extracting working condition characteristic value of marine main engine | |
| Yu et al. | Research of Fault Feature Extraction and Analysis Method Based on Aeroengine Fault Data |
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 |