CN111159646B - Grouping method for multi-working-condition performance data of fuel injector - Google Patents
Grouping method for multi-working-condition performance data of fuel injector Download PDFInfo
- Publication number
- CN111159646B CN111159646B CN201911341429.6A CN201911341429A CN111159646B CN 111159646 B CN111159646 B CN 111159646B CN 201911341429 A CN201911341429 A CN 201911341429A CN 111159646 B CN111159646 B CN 111159646B
- Authority
- CN
- China
- Prior art keywords
- data
- grouping
- main
- unqualified
- pca
- 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
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Analysis (AREA)
- Evolutionary Biology (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Mathematical Optimization (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Artificial Intelligence (AREA)
- Algebra (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Combined Controls Of Internal Combustion Engines (AREA)
- Fuel-Injection Apparatus (AREA)
Abstract
The invention belongs to the technical field of automobile power systems, and relates to a grouping method of multi-station performance data of an oil sprayer; the method comprises the following steps: 1. data cleaning and analyzing main components of the data; 2. a weight W is distributed to each main working condition; 3. summing the weighted working condition data; 4. sorting according to the summed values; 5. extracting corresponding number of oil injectors in sequence for grouping, and judging whether the grouping requirement is met; 6. if the data is not qualified, replacing the data according to the condition of the unqualified data, and recording the problem of the unqualified data; 7. if the data are qualified, continuing to execute the step 5 until all the data are extracted; 8. calculating grouping efficiency, if the grouping efficiency does not meet the set requirement, modifying the weight according to the unqualified record, and returning to the step 3 to continue execution; 9. if the set target is satisfied, the program is ended. All main characteristics of the data are comprehensively considered, the prejudice caused by focusing on only a single characteristic is avoided, a large amount of analysis time is saved, and the grouping efficiency is improved.
Description
Technical Field
The invention belongs to the technical field of automobile power systems, and relates to a grouping method of multi-station performance data of an oil sprayer.
Background
In order to make the injection of each cylinder of an engine uniform, it is required that injectors mounted on one engine be as uniform as possible. Therefore, before shipment, the injectors need to be grouped according to their performance data, as shown in fig. 1, each group contains 6 injectors (determined according to the number of cylinders of the engine). The grouping requirements are shown in Table 1.
TABLE 1 grouping requirement for fuel injectors
Because the performance consistency requirements of different engines on the fuel injectors under different working conditions are not consistent, specific working condition points are replaced by working conditions 1 to 6 respectively. The calculation formula of the fuel injection quantity deviation delta is shown in a formula (1), wherein Max, min, average is the maximum value, the minimum value and the average value of the group of data under a certain working condition respectively.
Current practice 1: the personnel operating the grouping are used for observing and analyzing all data needing to be grouped, such as partial data listed in the table 2, operating the EXCEL software, selecting out the data of 6 working conditions approximately similar, temporarily grouping the data into a group, calculating whether the data meet the grouping requirement in the table 1, and continuously searching for proper data if the data meet the grouping requirement.
Table 2 Multi-station performance data for fuel injector (cut-out data, 40 pieces)
Current practice 2: through programming, the performance data of the oil injectors are ordered according to a certain working condition, every 6 oil injectors are divided into a group according to the sequence, and finally whether the performance data are met or not is judged.
Disadvantages of approach 1: firstly, a group of fuel injector data is selected and analyzed each time, the EXCEL is manually operated to copy and paste, and under a large amount of data, the required time is long, and the labor cost is high. Secondly, the data can be estimated to be a group manually by experience, so that the randomness is high, the failure probability is high, and the time is wasted. And the probability of error of manual judgment is increased under the conditions of long-time work and large amount of data.
Disadvantages of approach 2: although the data is processed quickly by the program, a large amount of time can be saved compared with manual work, the algorithm of the program is too simple and is ordered according to a working condition. The method can only meet the data similarity of one working condition and can not ensure the requirements of other working conditions, so that the qualification rate is low, the grouping efficiency is low, and the grouping rate is far lower than the grouping rate of manual grouping.
Disclosure of Invention
The invention solves the technical problems of high manual operation time and labor cost, easy fatigue error, simple program grouping algorithm and low grouping efficiency in the prior art, and provides a grouping method of multi-station performance data of an oil sprayer.
In order to solve the technical problems, the invention is realized by adopting the following technical scheme, and the technical scheme is as follows in combination with the accompanying drawings:
a grouping method of multi-station performance data of an oil sprayer comprises the following steps:
1. data cleaning, analyzing main components of the data, selecting main features by utilizing PCA (principal component analysis), and selecting main working conditions according to the main features;
2. a weight W is distributed to each main working condition;
3. summing the weighted working condition data;
4. sorting according to the summed values;
5. extracting corresponding number of oil injectors in sequence for grouping, and judging whether the grouping requirement is met;
6. if the data is not qualified, replacing the data according to the condition of the unqualified data, and recording the problem of the unqualified data;
7. if the data are qualified, continuing to execute the step 5 until all the data are extracted;
8. calculating grouping efficiency, if the grouping efficiency does not meet the set requirement, modifying the weight according to the unqualified record, and returning to the step 3 to continue execution;
9. if the set target is satisfied, the program is ended.
According to the invention, the performance data of multiple working conditions of the oil injector are comprehensively considered, the oil injectors with close performance are divided into one group, the grouping efficiency is improved, and the grouping of the multi-working-condition performance data of the oil injector is realized.
The data in step 1 are cleaned, main components of the data are analyzed, main characteristics are selected by utilizing main component analysis PCA, and main working conditions are selected according to the main characteristics; the specific contents are as follows:
importing Data into a Data processing page of the EXCEL through a program, wherein each Data represents the Data of each working condition of one fuel injector, the columns represent different working conditions respectively, and Data represent all the Data; then the main component of all data is obtained by calling the python script; principal component analysis PCA in the Python data analysis package scikit-learn can calculate the principal component of a given data set; principal component analysis and PCA extraction are carried out to extract principal component, the feature number is reduced to 2 features, and the features are respectively marked as X 1 ，X 2 ；
Python script code implementation:
from sklearn.decomposition import PCA
pca=pca (0.95) # extraction principal component represents 95% of the original data
Data_new＝pca.fittransform(Data)。
In the step 2, a weight W is allocated to each main working condition; the specific contents are as follows:
each feature is initialized with a weight, denoted as W 1 ，W 2 Wherein 2 weights are required to satisfy a condition that their sum is 1; the first initialization may be arbitrarily assigned, for example, (0.5 ).
Step 3, summing the weighted working condition data; the specific calculation mode is as follows: calculate the weighted data D, d=x 1 *W 1 +X 2 *W 2 ；
Step 5, extracting corresponding number of oil injectors in sequence for grouping, and judging whether the grouping requirement is met; the specific contents are as follows:
sequentially extracting corresponding quantities to judge, wherein the specific quantities are determined by the cylinder numbers of the corresponding engines; judging whether the grouping specification requirements are met or not; to save the judgment logic, a large oil mass working condition and a small oil mass working condition are selected to replaceAll conditions, e.g. Q main ，Q small The method comprises the steps of carrying out a first treatment on the surface of the And according toCalculate Q main And Q small Is set in the fuel injection amount.
If the data is not qualified in the step 6, replacing the data according to the unqualified condition, and recording the unqualified problem; the specific contents are as follows:
if delta is out of tolerance, the minimum value is found, and the next line of data is replaced in sequence, so that the minimum value in the data is removed because the new data extracted later can only be larger.
The efficiency of grouping is calculated in the step 8, if the set requirement is not met, the weight is modified according to the unqualified record, and the step 3 is returned to be continuously executed; the specific contents are as follows:
and calculating whether all working conditions meet requirements, counting the grouping efficiency eta, eta=grouping number/(total number |6), and if the grouping efficiency eta is not up to the set requirements, modifying the weight according to the unqualified records, and returning to the step 3 to continue execution.
Compared with the prior art, the invention has the beneficial effects that:
the feature weighting algorithm is combined with the VBA program to realize the algorithm, all main features of the data are comprehensively considered, the prejudice caused by only focusing on a single feature is avoided, the grouping process adopts a similar pop method, the data are screened one by one to judge, and the whole group of data is prevented from being wasted due to large deviation of individual data. The invention not only saves a great amount of analysis time, but also has a grouping rate larger than that of manual processing and program processing of a simple ordering algorithm, and greatly improves grouping efficiency of grouping.
Drawings
The invention is further described below with reference to the accompanying drawings:
FIG. 1 is a schematic diagram of a fuel injector grouping;
FIG. 2 is a flow chart of a fuel injector grouping;
FIG. 3 is a user interface for data cleansing;
FIG. 4 is a user interface for fuel injector grouping.
Detailed Description
The invention is described in detail below with reference to the attached drawing figures:
the invention adopts VBA program to realize algorithm, designs window program, and is convenient for operators to directly operate EXCEL to process data. The program comprises the functions of data cleaning, data grouping, data judging, data statistics, manual adjustment and the like.
1. Referring to fig. 3, first, data is imported into the EXCEL Data processing page by a program, each Data represents Data of each working condition of one fuel injector, columns represent different working conditions, and Data represents all the Data. And then the main component of all data is obtained by calling the python script. PCA (principal component analysis) in the Python data analysis package scikit-learn can calculate the principal component of a given data set; in the data in this example, principal components are extracted by PCA (principal component analysis), and the number of original features is reduced to 2 features. Respectively marked as X 1 ，X 2 ；
Python script code implementation:
from sklearn.decomposition import PCA
pca=pca (0.95) # extracted principal component represents 95% of the original data.
Data_new＝pca.fittransform(Data)
2. Referring to FIG. 2, a weight is initialized for each feature, denoted as W 1 ，W 2 Wherein 2 weights need to satisfy a condition that their sum is 1. The first initialization can be arbitrarily assigned, for example, (0.5 );
3. calculate the weighted data D, d=x 1 *W 1 +X 2 *W 2 ；
4. Because the data of the original fuel injector is in one-to-one correspondence with the data column of the D, after all the data are ordered according to the weighted data D, the original data are ordered accordingly;
5. the corresponding number is extracted sequentially, in this example, 6 fuel injectors are extracted to judge whether the fuel injectors meet the standard requirement of grouping according to the table 1, and in order to save judgment logic, the fuel injectors are selectedA large oil condition and a small oil condition instead of all conditions, e.g. Q main ，Q small . And calculate Q according to equation (1) main And Q small The actual fuel injection quantity deviation delta;
6. if the data is disqualified, the minimum value is found out, and the data of the next row is replaced in sequence (the minimum value in the data is removed because the new data extracted later can only be larger, the new data added is combined with the original 5 pieces of left data, so that the new data is possibly more consistent, otherwise, the situation is worse);
7. referring to fig. 1, if the specification requirements are met, the 6 lines of data (6 injector data) are grouped and placed in the completion zone (other page). The program continues to execute the step 5 until all data are extracted;
8. referring to fig. 4, it is calculated whether all the working conditions meet the requirements, because only 2 working conditions are considered in the judging process of step 6, although the 2 working conditions are qualified through weighted sequencing, the other working conditions cannot be completely guaranteed to be qualified, and W 1 ，W 2 And the solution is not necessarily the optimal solution, so that the efficiency eta of the grouping is counted, eta=the grouping number/(the total number |6), and if the set requirement is not met, the weight is modified according to the unqualified record, and the step 3 is returned to be continuously executed. Analyzing records which cannot be grouped, and increasing W if the number of unqualified large-oil-amount working conditions is large 1 If the quantity of unqualified small oil quantity working conditions is more, the W is increased 2 ；
The symbol "|" represents integer division.
9. If the set target is satisfied, the program is ended.
The invention is practically applied to the grouping of the fuel injectors of the common rail system of the engine in the subdivision field, and can also be applied to the grouping of multi-station data or multi-feature data according to the given grouping specification in other fields.
Claims (4)
1. A method for grouping multiple operation performance data of an oil injector, comprising the steps of:
step 1: data cleaning, analyzing main components of the data, analyzing PCA (principal component analysis) to select main features, and selecting main working conditions according to the main features;
step 2: a weight W is distributed to each main working condition;
in the step 2, a weight W is allocated to each main working condition; the specific contents are as follows:
each feature is initialized with a weight, denoted as W 1 ，W 2 Wherein 2 weights are required to satisfy a condition that their sum is 1; the first initialization can be arbitrarily assigned;
step 3: summing the weighted working condition data;
step 3, summing the weighted working condition data; the specific calculation mode is as follows: calculate the weighted data D, d=x 1 *W 1 +X 2 *W 2 ；
Step 4: sorting according to the summed values;
step 5: extracting corresponding number of oil injectors in sequence for grouping, and judging whether the grouping requirement is met;
step 5, extracting corresponding number of oil injectors in sequence for grouping, and judging whether the grouping requirement is met; the specific contents are as follows:
sequentially extracting corresponding quantities to judge, wherein the specific quantities are determined by the cylinder numbers of the corresponding engines; judging whether the grouping specification requirements are met or not; to save the judgment logic, a large oil mass working condition and a small oil mass working condition are selected to replace all working conditions, such as Q main ，Q small The method comprises the steps of carrying out a first treatment on the surface of the And according toCalculate Q main And Q small The actual fuel injection quantity deviation delta;
step 6: if the data is not qualified, replacing the data according to the condition of the unqualified data, and recording the problem of the unqualified data;
step 7: if the data are qualified, continuing to execute the step 5 until all the data are extracted;
step 8: calculating grouping efficiency, if the grouping efficiency does not meet the set requirement, modifying the weight according to the unqualified record, and returning to the step 3 to continue execution;
step 9: if the set target is satisfied, the program is ended.
2. A method of grouping multi-event performance data for a fuel injector as set forth in claim 1 wherein:
the data in step 1 are cleaned, main components of the data are analyzed, main characteristics are selected by utilizing main component analysis PCA, and main working conditions are selected according to the main characteristics; the specific contents are as follows:
importing Data into a Data processing page of the EXCEL through a program, wherein each Data represents the Data of each working condition of one fuel injector, the columns represent different working conditions respectively, and Data represent all the Data; then the main component of all data is obtained by calling the python script; principal component analysis PCA in the Python data analysis package scikit-learn can calculate the principal component of a given data set; principal component analysis and PCA extraction are carried out to extract principal component, the feature number is reduced to 2 features, and the features are respectively marked as X 1 ，X 2 ；
Python script code implementation:
from sklearn.decomposition import PCA；
pca=pca (0.95) # extraction principal component represents 95% of the original data;
Data_new＝pca.fittransform(Data)。
3. a method of grouping multi-event performance data for a fuel injector as claimed in claim 2, wherein:
if the data is not qualified in the step 6, replacing the data according to the unqualified condition, and recording the unqualified problem; the specific contents are as follows:
if delta is out of tolerance, the minimum value is found, and the next line of data is replaced in sequence, so that the minimum value in the data is removed because the new data extracted later can only be larger.
4. A method of grouping multi-event performance data for a fuel injector as claimed in claim 3, wherein:
the efficiency of grouping is calculated in the step 8, if the set requirement is not met, the weight is modified according to the unqualified record, and the step 3 is returned to be continuously executed; the specific contents are as follows:
calculating whether all working conditions meet requirements, counting the grouping efficiency eta, eta=grouping number/(total number=6), and if the grouping efficiency eta is not up to the set requirements, modifying the weight according to unqualified records, and returning to the step 3 to continue execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911341429.6A CN111159646B (en) | 2019-12-24 | 2019-12-24 | Grouping method for multi-working-condition performance data of fuel injector |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911341429.6A CN111159646B (en) | 2019-12-24 | 2019-12-24 | Grouping method for multi-working-condition performance data of fuel injector |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111159646A CN111159646A (en) | 2020-05-15 |
| CN111159646B true CN111159646B (en) | 2023-08-15 |
Family
ID=70558219
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911341429.6A Active CN111159646B (en) | 2019-12-24 | 2019-12-24 | Grouping method for multi-working-condition performance data of fuel injector |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111159646B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101178773A (en) * | 2007-12-13 | 2008-05-14 | 北京中星微电子有限公司 | Image recognition system and method based on characteristic extracting and categorizer |
| CN103390154A (en) * | 2013-07-31 | 2013-11-13 | 中国人民解放军国防科学技术大学 | Face recognition method based on extraction of multiple evolution features |
| CN103488790A (en) * | 2013-10-08 | 2014-01-01 | 河海大学 | Polychronic time sequence similarity analysis method based on weighting BORDA counting method |
| CN105528516A (en) * | 2015-12-01 | 2016-04-27 | 三门县人民医院 | Clinic pathology data classification method based on combination of principal component analysis and extreme learning machine |
| CN107798696A (en) * | 2017-10-19 | 2018-03-13 | 北京联合大学 | A kind of three-dimensional point cloud method for registering based on guarantor office PCA |
| CN109815990A (en) * | 2018-12-28 | 2019-05-28 | 天津大学 | Principal component analysis system based on centralization weight |
| CN110021386A (en) * | 2017-07-27 | 2019-07-16 | 北大医疗信息技术有限公司 | Feature extracting method and feature deriving means, equipment, storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103077402B (en) * | 2012-12-28 | 2016-05-11 | 国家电网公司 | Partial discharge of transformer mode identification method based on singular value decomposition algorithm |
-
2019
- 2019-12-24 CN CN201911341429.6A patent/CN111159646B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101178773A (en) * | 2007-12-13 | 2008-05-14 | 北京中星微电子有限公司 | Image recognition system and method based on characteristic extracting and categorizer |
| CN103390154A (en) * | 2013-07-31 | 2013-11-13 | 中国人民解放军国防科学技术大学 | Face recognition method based on extraction of multiple evolution features |
| CN103488790A (en) * | 2013-10-08 | 2014-01-01 | 河海大学 | Polychronic time sequence similarity analysis method based on weighting BORDA counting method |
| CN105528516A (en) * | 2015-12-01 | 2016-04-27 | 三门县人民医院 | Clinic pathology data classification method based on combination of principal component analysis and extreme learning machine |
| CN110021386A (en) * | 2017-07-27 | 2019-07-16 | 北大医疗信息技术有限公司 | Feature extracting method and feature deriving means, equipment, storage medium |
| CN107798696A (en) * | 2017-10-19 | 2018-03-13 | 北京联合大学 | A kind of three-dimensional point cloud method for registering based on guarantor office PCA |
| CN109815990A (en) * | 2018-12-28 | 2019-05-28 | 天津大学 | Principal component analysis system based on centralization weight |
Non-Patent Citations (1)
| Title |
|---|
| "电控汽油喷射发动机的自诊断系统（下）";李潍;《北京汽车》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111159646A (en) | 2020-05-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106599052B (en) | Apache Kylin-based data query system and method | |
| CN108182063A (en) | A kind of implementation method of big data analysis visual configuration | |
| CN111159646B (en) | Grouping method for multi-working-condition performance data of fuel injector | |
| CN110263291B (en) | Industrial data trend identification method and system | |
| CN103455509B (en) | A kind of method and system obtaining time window model parameter | |
| CN112182568B (en) | Malicious code classification based on graph convolution network and topic model | |
| CN116187650A (en) | Autonomous machine learning demand prediction method and system for product parts | |
| CN117455222B (en) | Solving method based on distributed heterogeneous flow shop group scheduling problem | |
| CN110569237A (en) | System and method for realizing real-time data cleaning processing | |
| CN114238360A (en) | User behavior analysis system | |
| CN110516123A (en) | Data query time-consuming appraisal procedure, device, computer equipment and storage medium | |
| CN109470954A (en) | A kind of operation of power networks condition monitoring system and its monitoring method based on big data | |
| CN106790258A (en) | A kind of method and system of screening server network request | |
| CN114358924A (en) | Method and device for screening limit adjustment strategy, computer equipment and storage medium | |
| CN114862055A (en) | Multi-source data driving-based blast furnace gas prediction method for iron and steel enterprise | |
| KR101632537B1 (en) | Technical ripple effect analysis method | |
| CN111857192A (en) | Method and device for regulating ejector | |
| CN112329013A (en) | Malicious code classification method based on graph convolution network and topic model | |
| Sanderson et al. | Maintaining performance with less data | |
| CN1549574A (en) | Telephone bill data processing method and apparatus | |
| CN117130791B (en) | Computing power resource allocation method and system of cloud customer service platform | |
| CN109992329A (en) | A kind of method grade method for analyzing performance based on rear end grain technology | |
| CN113011831B (en) | System for data auditing and data auditing method | |
| CN108280126A (en) | A kind of big data analysis processing method | |
| CN110298588B (en) | Automatic standard optimization system and method based on statistical analysis |
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 |