Diesel engine combustion development calibration method and device based on principal component analysis

Abstract

The invention discloses a diesel engine combustion development calibration method and device based on principal component analysis. The method comprises the following steps that a test data set of an engine is obtained, and parameters in the test data set are normalized; bringing into a principal component analysis model according to the normalized experimental data set, and constructing a covariance matrix of each index parameter in the principal component analysis model; solving a characteristic value and a characteristic vector of the covariance matrix; determining the contribution rate of each principal component factor in a principal component analysis model according to the feature values and the feature vectors; screening the principal component factors according to the contribution rate to obtain dimensionality reduction data less than the dimensionality of the index data; calculating the ratio of the sum of the product of each dimensionality reduction data and the corresponding contribution rate to the sum of the contribution rates of all dimensionality reduction data; and screening out the minimum value in the ratios. According to the method, rapid calibration is realized by performing dimension reduction on data, and the repetition amount and complexity of calibration work are reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of combustion development calibration, and in particular relates to a diesel engine combustion development calibration method and equipment based on principal component analysis. Background Art

[0002] At present, there are many indicators that need to be paid attention to in diesel engine development, such as fuel consumption, exhaust temperature, explosion pressure, various gas pollutants, smoke density, etc. Combustion development is to adjust the engine combustion parameters to make the indicators reach a suitable value. However, when adjusting the diesel engine combustion parameters, most of the indicator values ​​are contradictory, that is, adjusting the diesel engine combustion parameters will make some indicators better and other indicators worse. After the diesel engine test, the data between the indicators corresponding to many different combustion parameters will be recorded. It is bound to be time-consuming and labor-intensive to select a suitable set of indicators from many combinations, and manual iterative adjustments are required. Usually, during the test, 150 data between the indicators corresponding to different combustion parameters will be recorded. From this data, a suitable indicator is manually selected. Nine dimensional indicators need to be considered when selecting: specific fuel consumption BFSC, nitrogen oxides NO E, smoke density SOOT, oil volume InjCtl_qSetUnBal, specific nitrogen oxide EM_NOX_1, specific hydrocarbon EM_THC, explosion pressure P_MAX, exhaust temperature T_Exh, turbocharger speed turbo_speed, among which, specific fuel consumption represents the amount of diesel consumed when the engine outputs 1Kw·h of mechanical work, and the unit is g / Kw·h; nitrogen oxide represents the nitrogen oxide emission concentration, and the unit is g / Kw·h; smoke density represents the concentration of solid particulate matter emitted by the diesel engine, and the unit is g / h; oil volume represents the fuel consumption per stroke efficiency of the engine, and the unit is mg / hub; specific nitrogen oxide emission concentration is in ppm; specific hydrocarbon emission concentration is in ppm; explosion pressure represents the maximum pressure value in the combustion chamber during the engine operation stage, and the unit is bar; exhaust temperature is in ℃; turbocharger speed is in rpm.

[0003] In the above indicators, each indicator is required to be as small as possible. Due to the intricate and complex situation of the indicators, the indicator dimension is 9, and the workload of selecting 1 out of 150, the selection process is time-consuming and laborious, and the accuracy of the final indicator also needs to be repeatedly confirmed. This work has become the key difficulty of combustion development. Summary of the invention

[0004] In order to solve the problem that the existing calibration methods are time-consuming and labor-intensive, the present invention provides a diesel engine combustion development calibration method and equipment based on principal component analysis, which realizes rapid calibration by reducing the dimension of data, thereby reducing the duplication and complexity of calibration work.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] A first aspect of the present invention provides a diesel engine combustion development calibration method based on principal component analysis, comprising the following steps:

[0007] Acquire a test data set of the engine and normalize parameters in the test data set, wherein the test data set includes at least one set of combustion parameters and index parameters corresponding thereto, wherein the index parameters include specific fuel consumption, nitrogen oxides, smoke density, oil volume, specific nitrogen oxides, specific hydrocarbons, explosion pressure, exhaust temperature, and supercharger speed;

[0008] According to the normalized experimental data set, the principal component analysis model is introduced and the covariance matrix of each indicator parameter in the principal component analysis model is constructed;

[0009] Solve for the eigenvalues ​​and eigenvectors of the covariance matrix;

[0010] Determine the contribution rate of each principal component factor in the principal component analysis model according to the eigenvalue and the eigenvector;

[0011] Screening the principal component factors according to the contribution rate to obtain dimension-reduced data having a number of dimensions less than the indicator data;

[0012] Calculate the ratio of the sum of the products of each dimension-reduced data and the corresponding contribution rate to the sum of the contribution rates of all dimension-reduced data;

[0013] The minimum value among the ratios is selected.

[0014] The second aspect of the present invention provides a diesel engine combustion development calibration device based on principal component analysis, comprising a memory and a controller that are communicatively connected in sequence, wherein a computer program is stored in the memory, and the controller is used to read the computer program and execute a diesel engine combustion development calibration method based on principal component analysis described in the first aspect.

[0015] Compared with the prior art, the present invention has at least the following advantages and beneficial effects:

[0016] The present invention reduces the dimensionality of index data during diesel engine combustion development and reduces the index to one-dimensional data through contribution rate, thereby reducing test development costs and resource occupation and reducing the duplication and complexity of calibration work. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0018] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0019] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0020] It should be noted that:

[0021] The terms “first”, “second”, etc. are only used for distinguishing descriptions and should not be understood as indicating or implying relative importance.

[0022] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0023] The present invention discloses a diesel engine combustion development calibration method based on principal component analysis, which comprises steps S01 to S06.

[0024] Step S01, obtaining a test data set of the engine and normalizing the parameters in the test data set, wherein the test data set includes at least one set of combustion parameters and index parameters corresponding thereto, wherein the index parameters include specific fuel consumption, nitrogen oxides, smoke density, oil volume, specific nitrogen oxides, specific hydrocarbons, explosion pressure, exhaust temperature, and supercharger speed.

[0025] The accuracy of the data source is directly related to the accuracy of the calibration results. In this regard, in this step, the obtained data set needs to be preprocessed, that is, the data with singular points and missing values ​​are deleted.

[0026] Specifically, the preprocessing adopts the following method: first calculate the mean value and standard deviation of the data of each dimension of the indicator parameter; then take three times the mean value as the center and the standard deviation as the radius, filter out the data within the range and delete the missing values.

[0027] The parameter normalization in the experimental dataset includes:

[0028]

[0029] Among them, x ij is the parameter in the test data set, y ij is the normalized value.

[0030] Since the indicator parameter has 9 dimensions, it is necessary to normalize the 9-dimensional data to obtain a new matrix y. The mean of the processed data is 0, the variance is 1, and each element in the matrix y is y ij .

[0031] Step S02: Bring the normalized experimental data set into the principal component analysis model and construct the covariance matrix of each indicator parameter in the principal component analysis model.

[0032] In this step, the principal component analysis model is:

[0033]

[0034] y is the ij The matrix is ​​composed of: P is 9, Y1, Y2, ...Y p are the normalized row vectors of the j-th index parameters, and j is 1, 2, 3…p.

[0035] Principal component analysis attempts to define a set of uncorrelated variables, called Y1, Y2, ...Y p The principal components of p , each principal component is Y1, Y2, ...Y p A linear combination of .

[0036] Linear combination is essentially the inner product of vectors, solving the principal components Z1, Z2, ...Z p That is, solve a1, a2, ...a p The sample covariance matrix can be used to solve a1, a2, ...a p .

[0037] The covariance matrix is:

[0038]

[0039] in,

[0040]

[0041] k is 1, 2, 3…p.

[0042] Step S03, solving the eigenvalues ​​and eigenvectors of the covariance matrix.

[0043] Step S04: determining the contribution rate of each principal component factor in the principal component analysis model according to the eigenvalue and the eigenvector.

[0044] In this step, the principal component factor Z is first determined j , and then according to the eigenvalue λ jThe contribution rate X of the principal component factor is calculated by eigenvector.

[0045] in,

[0046] Z j =e′ j y=e j1 Y1+e j2 Y2+…+e jp Y p ,

[0047]

[0048] e1, e2, ...e p is the standardized eigenvector, and its corresponding eigenvalues ​​are: λ1, λ2, ...λ p .

[0049] λ1≥λ2≥…≥λ p ≥0.

[0050] Step S05: screening the principal component factors according to the contribution rates to obtain dimension-reduced data having a number of dimensions less than that of the indicator data.

[0051] Generally, when the cumulative contribution rate is between 85% and 95%, it indicates that the selected principal component factors can represent most of the characteristics of the original variables. When screening, the cumulative screening is carried out in descending order of contribution rate. When the contribution rate is between 85% and 95%, it can be determined that the reduced dimensionality data with less dimensions than the indicator data is screened.

[0052] Step S06, calculating the ratio of the sum of the product of each dimension-reduced data and the corresponding contribution rate to the sum of the contribution rates of all dimension-reduced data; and screening out the minimum value of the ratio.

[0053] In order to facilitate the understanding of the method of this solution, an example is given.

[0054] The combustion development calibration of the YN490plus diesel engine was performed, and 150 rows of data were collected. Each row contains 9 dimensional indicator data: specific fuel consumption BFSC, nitrogen oxides NOE, smoke density SOOT, oil volume InjCtl_qSetUnBal, specific nitrogen oxides EM_NOX_1, specific hydrocarbons EM_THC, explosion pressure P_MAX, exhaust temperature T_Exh, and turbocharger speed turbo_speed. Some of the collected data are shown in Table 1.

[0055] Table 1: Diesel engine combustion development test data

[0056]

[0057]

[0058] Next, the data with singular points and missing values ​​are deleted, the index data are normalized, and the principal component analysis model is used to reduce the dimensionality of the experimental data.

[0059] After normalizing the 9-dimensional index data, the covariance of the normalized data is calculated, and the covariance matrix S is obtained as shown in Table 2.

[0060] Table 2

[0061] 1 -0.352 0.195 0.991 -0.295 0.203 -0.287 0.778 0.902 -0.352 1 -0.271 -0.311 0.968 0.019 0.926 -0.516 -0.225 0.195 -0.271 1 0.222 -0.178 -0.471 -0.375 0.306 0.162 0.991 -0.311 0.222 1 -0.266 0.201 -0.237 0.745 0.912 -0.295 0.968 -0.178 -0.266 1 -0.09 0.823 -0.353 -0.251 0.203 0.019 -0.471 0.201 -0.09 1 0.116 0.085 0.153 -0.287 0.926 -0.375 -0.237 0.823 0.116 1 -0.632 -0.041 0.778 -0.516 0.306 0.745 -0.353 0.085 -0.632 1 0.462 0.902 -0.225 0.162 0.912 -0.251 0.153 -0.041 0.462 1

[0062] The calculated covariance matrix eigenvalues ​​are shown in Table 3.

[0063] Table 3

[0064] <![CDATA[λ1]]> <![CDATA[λ2]]> <![CDATA[λ3]]> <![CDATA[λ4]]> <![CDATA[λ5]]> <![CDATA[λ6]]> <![CDATA[λ7]]> <![CDATA[λ8]]> <![CDATA[λ9]]> 4.374 2.197 1.391 0.577 0.42 0.021 0.015 0.004 0.001

[0065] The corresponding standardized feature vectors are shown in Table 4.

[0066] Table 4

[0067] <![CDATA[e1]]> <![CDATA[e2]]> <![CDATA[e3]]> <![CDATA[e4]]> <![CDATA[e5]]> <![CDATA[e6]]> <![CDATA[e7]]> <![CDATA[e8]]> <![CDATA[e9]]> 0.409 0.34 0.056 0.012 0.105 0.308 0.091 0.506 -0.59 -0.37 0.391 0.195 0.172 -0.05 0.191 0.059 -0.63 -0.45 0.182 -0.2 0.628 0.032 -0.73 -0.04 -0.01 0.05 -0.04 0.399 0.361 0.083 -0.02 0.035 0.244 -0.7 -0.22 0.331 -0.33 0.364 0.298 0.402 0.074 0.255 0.239 0.354 0.503 0.043 0.273 -0.671 0.23 -0.64 0.035 0.083 0.023 0.066 -0.34 0.442 0.077 -0.25 -0.08 -0.63 -0.33 0.301 -0.12 0.401 0.059 0.067 0.67 0.195 -0.56 0.089 -0.15 -0.04 0.336 0.397 0.088 -0.5 0.004 -0.16 0.569 -0.24 0.261

[0068] The contribution rates of the principal component factors are shown in Table 5.

[0069] Table 5

[0070]

[0071] Here, taking 3 principal components will reach a cumulative contribution rate of 88.5%. Now, substituting the eigenvectors into the principal component matrix, we can get:

[0072]

[0073] The calculation results are shown in Table 6.

[0074] Table 6: Principal component analysis dimensionality reduction table

[0075] Table 6

[0076]

[0077] Using the factor contribution rate as the weighting coefficient, the 3D data is reduced to 1D data, as shown in Table 7:

[0078] Table 7

[0079]

[0080] By sorting the 1-dimensional data in Table 7, the optimal result can be obtained.

[0081] Among the optimal results, BFSC is 216.4 g / Kw·h, NOE is 5.1 g / Kw·h, SOOT is 0.4 g / h, injctl_qsetunbal is 21.7 mg / hub, EM_NOX is 376.8 ppm, EM_THC is 63.1 ppm, P_MAX is 60.9 bar, T_Exh_2 is 301.6℃, turbo_speed is 101052.8 rpm, and the derived value is -0.482, which is the smallest row among 150 rows.

[0082] Compared with the results of manual selection based on experience, the recommended data after dimensionality reduction through principal component analysis can meet the requirements of this operating condition. Manual selection of the optimal value from 150 records is time-consuming and inefficient. In contrast, the present invention can standardize the combustion development and calibration process of diesel generators by establishing a principal component analysis model, which is more objective and effective than the previous method based on experience, simplifies the entire workflow, reduces experimental development costs and resource usage, and reduces the duplication and complexity of calibration work.

[0083] The third aspect of the present invention provides a diesel engine combustion development calibration device based on principal component analysis, including a memory and a controller connected in sequence, wherein a computer program is stored on the memory, and the controller is used to read the computer program and execute a diesel engine combustion development calibration method based on principal component analysis described in the first aspect and any of its possibilities. Specifically, the memory may include, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a flash memory, a first-input first-output memory (FIFO), and / or a first-input last-output memory (FILO), etc.; the controller may not be limited to a microcontroller of the STM32F105 series. In addition, the computer device may also include, but is not limited to, a power supply unit, a display screen, and other necessary components.

[0084] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A diesel engine combustion development calibration method based on principal component analysis, characterized in that: The following steps are involved: Acquire an engine test data set and normalize the parameters in the test data set, wherein the test data set includes at least one set of combustion parameters and index parameters corresponding thereto, wherein the index parameters include specific fuel consumption, nitrogen oxides, smoke density, oil volume, specific nitrogen oxides, specific hydrocarbons, explosion pressure, exhaust temperature, and supercharger speed; bring the normalized experimental data set into a principal component analysis model and construct a covariance matrix of each index parameter in the principal component analysis model; Solve for the eigenvalues ​​and eigenvectors of the covariance matrix; Determine the contribution rate of each principal component factor in the principal component analysis model according to the eigenvalue and the eigenvector; Screening the principal component factors according to the contribution rate to obtain dimension-reduced data having a number of dimensions less than the indicator data; Calculate the ratio of the sum of the products of each dimension-reduced data and the corresponding contribution rate to the sum of the contribution rates of all dimension-reduced data; and filter out the minimum value among the ratios.

2. The diesel engine combustion development calibration method based on principal component analysis according to claim 1 is characterized in that: After obtaining the test data set of the engine, and before normalizing the test data set, the method further includes: Delete data with singular points and missing values.

3. The diesel engine combustion development calibration method based on principal component analysis according to claim 2 is characterized in that: The data for deleting singular points and missing values ​​includes: Calculate the mean and standard deviation of each dimension of the indicator parameter; With three times of the mean as the center and the standard deviation as the radius, the data within the range are screened out and missing values ​​are deleted.

4. The diesel engine combustion development calibration method based on principal component analysis according to claim 1, characterized in that: The parameter normalization in the test data set includes: Among them, x ij is the parameter in the test data set, y ij is the normalized value.

5. The diesel engine combustion development calibration method based on principal component analysis according to claim 1, characterized in that: The principal component analysis model is: y is the ij The matrix is ​​composed of: P is 9, Y1, Y2, ...Y p are the normalized row vectors of the j-th index parameter, j is 1, 2, 3…p; The covariance matrix is: in, k is 1, 2, 3…p.

6. The diesel engine combustion development calibration method based on principal component analysis according to claim 1, characterized in that: Determining the contribution rate of each principal component factor in the principal component analysis model according to the eigenvalue and the eigenvector includes: Determine the principal component factor Z j , Z j =e′ j y=e j1 Y1+e j2 Y2+…+e jp Y p , Among them, e1, e2, ... ep 是 Normalized eigenvectors, whose corresponding eigenvalues ​​are: λ1, λ2, ...λ p ; According to the eigenvalue λ j The contribution rate X of the principal component factor is calculated by the eigenvector.

7. A diesel engine combustion development calibration device based on principal component analysis, comprising a memory and a controller connected in sequence, wherein a computer program is stored in the memory, characterized in that: The controller is used to read the computer program and execute a diesel engine combustion development calibration method based on principal component analysis as described in any one of claims 1 to 6.

Images