Engine flameout problem classification method and device and storage medium
Through the method of principal component analysis and support vector machine, the vehicle parameters are collected and reduced in dimensionality, and a separate hyperplanar model is constructed, which solves the problem of occasional fire stall caused by brakes by unmanned vehicles in mines, and achieves efficient fire stall prediction and prevention.
Patent Information
- Application Number
- CN202510566393.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-30
- Publication Date
- 2025-08-19
AI Technical Summary
In the mine unmanned driving scenario, the occasional fire stalling problem of automatic transmission vehicles when the brake force is too high or too fast is difficult to predict and prevent, affecting production efficiency and bringing safety risks.
The principal component analysis and support vector machine method are used to construct a sample set by collecting vehicle parameters, dimensionality reduction processing and binary classification training, and a separate hyperplanar model is generated to achieve the prediction of engine shutdown.
It has achieved high accuracy prediction of the problem of unmanned vehicles in mines with low false alarm rate, which can prevent the risk of fire out in advance and improve production safety.
Smart Images
Figure CN120508892A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of unmanned driving technology in mines, and in particular to a method, device and storage medium for classifying engine stall problems based on principal component analysis and support vector machines. Background Art
[0002] With the rapid development of science and technology, autonomous driving technology has made significant progress. Developers are increasingly eager to put autonomous driving technology into practical use. Mining scenarios, with their relatively closed environments and fixed driving routes, have become one of the most suitable scenarios for the implementation of autonomous driving technology.
[0003] In the field of unmanned mining, rigid fuel vehicles still account for a large proportion, and most of them use automatic transmissions. However, in actual use, automatic transmissions have a problem that cannot be ignored: when the braking force is too strong or too fast, the vehicle is likely to stall. It is worth noting that this stalling problem does not necessarily occur every time the braking force is strong, but rather exhibits a certain degree of sporadic occurrence, which makes troubleshooting and prevention more difficult. Once the stalling problem occurs, it will not only affect the normal operation of mining operations and reduce production efficiency, but also may bring potential safety risks. For example, on some slopes, the vehicle may roll away after stalling.
[0004] Currently, there is a lack of effective prediction and prevention measures for this problem. It is impossible to accurately predict the stalling of the automatic transmission during braking, and it is difficult to take measures in advance through technical means to prevent the occurrence of stalling problems.
[0005] It can be seen that there is an urgent need to develop an effective method that can predict whether the engine will stall under the current state. Summary of the Invention
[0006] In a first aspect of the present invention, in order to solve the above technical problems, the present invention provides a method for classifying engine stall problems, the method comprising: Collecting vehicle parameters with and without the engine turned off to form an original sample set; decentralizing the original sample set to obtain a decentralized sample matrix; wherein the vehicle parameters include vehicle speed, gear position, brake pressure, engine speed, retarder status, vehicle load, and current slope; Constructing a covariance matrix based on the centered sample matrix and calculating all eigenvalues and corresponding eigenvectors; Determine the dimension after dimensionality reduction according to the cumulative percentage of the eigenvalues, construct a transformation matrix and perform dimensionality reduction processing on the samples to generate a sample data set after dimensionality reduction; Based on the sample data set after dimensionality reduction, a support vector machine is used to perform binary classification training to obtain a separating hyperplane and a binary classification model for the engine stall problem; The vehicle parameters collected in real time are classified and predicted based on the binary classification model to determine whether the engine is at risk of stalling.
[0007] Furthermore, the vehicle parameters collected when the engine is turned off and when the engine is not turned off constitute an original sample set, including: Define a vector ,in, is the vehicle speed, For the gear position, is the brake pressure, is the engine speed, is the retarder state, is the vehicle load, is the current slope; The collected The data when the engine is turned off is expressed as ; The collected The data when the engine is not turned off is expressed as ; but The original sample set composed of group data is .
[0008] Furthermore, the decentralizing process of the original sample set to obtain a decentralized sample matrix specifically includes: Calculate the mean vector of the original sample set : ; Subtract the mean vector from each sample to obtain the decentralized sample : ; According to the sample , and obtain the decentralized sample matrix ,in .
[0009] Furthermore, the covariance matrix is constructed based on the centralized sample matrix, and all eigenvalues and corresponding eigenvectors are calculated, including: Define the covariance matrix : ; Using the Matlab function eig( ) calculated eigenvalues sorted from largest to smallest and the corresponding feature vectors .
[0010] Furthermore, determining the dimension after dimensionality reduction according to the cumulative percentage of the eigenvalues includes: Calculate the sum of the absolute values of all features : , where for The eigenvalue of the sequence number, is the number of features; Starting from the first of the mentioned eigenvalues, calculate the cumulative percentage : ; until the cumulative percentage is not less than 99%, the corresponding number of eigenvalues is the dimension after dimensionality reduction.
[0011] Furthermore, the binary classification training using a support vector machine includes: Construct a quadratic programming problem under soft constraint parameters and use the quadratic programming function quadprog provided by Matlab to solve and obtain the optimal weight vector and bias term; Based on the optimal weight vector and the bias term, the separating hyperplane and the binary classification model are determined, wherein the binary classification model satisfies the expression:
[0012] Where, is the optimal weight vector, is the bias term.
[0013] Furthermore, determining whether the engine is at risk of stalling includes: like =-1, the binary classification model predicts that the engine is not turned off and the vehicle continues to run as usual; like =1, the binary classification model predicts a flameout risk, triggering the vehicle control module to adjust the brake command to avoid the flameout risk.
[0014] Furthermore, the method is applied to rigid fuel mining vehicles in unmanned mine driving scenarios to predict engine stalling problems caused by braking operations of the vehicle.
[0015] According to a second aspect of the present invention, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the engine stall problem classification method when executing the computer program.
[0016] According to a third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the engine stall problem classification method are implemented.
[0017] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: Aiming at the problem of engine stalling of unmanned vehicles in mines, the present invention adopts the principal component analysis method to reduce the data dimension based on several pre-selected variables related to the engine stalling problem, and reduces the original sample matrix into a new sample matrix with a smaller dimension; based on the new sample matrix, the support vector machine method is used to find the separating hyperplane and obtain a binary classification model, thereby realizing the prediction of the stalling problem. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0019] Figure 1 This is a flowchart of the present invention. DETAILED DESCRIPTION
[0020] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.
[0021] The present invention aims to provide an engine stall problem classification method based on principal component analysis and support vector machine. The method is applied to rigid fuel mining vehicles in unmanned mine driving scenarios to predict engine stall problems caused by vehicle braking operations.
[0022] See also Figure 1 , specifically including the following steps: S1. Collect vehicle parameters when the engine is turned off and when it is not turned off to form an original sample set.
[0023] In a further solution, when analyzing the engine stall problem, the following vehicle parameters that may be related to the problem are selected based on empirical methods, including vehicle speed, gear position, brake pressure, engine speed, retarder status, vehicle load and current slope.
[0024] Define a vector ,in, is the vehicle speed, For gear, is the brake pressure, is the engine speed, Retarder status, For vehicle load, is the current slope.
[0025] Through data collection, we can obtain Data when the engine is turned off and Data when the engine is not turned off ;constitute samples , which is the original sample set.
[0026] S2. Decentralize the original sample set to obtain a decentralized sample matrix.
[0027] In the further scheme, first, the mean vector of the original sample set is calculated : .
[0028] Then, the mean vector is subtracted from each sample to obtain the decentralized sample : .
[0029] Then, based on the decentralized samples , and obtain the decentralized sample matrix : ,in .
[0030] S3. Construct a covariance matrix based on the centered sample matrix and calculate all eigenvalues and corresponding eigenvectors.
[0031] In the further scheme, first, define the covariance matrix : ; Then, use the matlab function eig( ) calculated eigenvalues sorted from largest to smallest and the corresponding feature vectors .
[0032] S4. Determine the dimension after dimensionality reduction based on the cumulative percentage of the eigenvalues, construct a transformation matrix and perform dimensionality reduction on the samples to generate a sample data set after dimensionality reduction.
[0033] Preferably, the data after dimensionality reduction is required to retain no less than 99% of the accuracy of the original information.
[0034] S41. Calculate the sum of the absolute values of all features : , where for The eigenvalue of the sequence number, is the number of features.
[0035] S42. Starting from the first eigenvalue, calculate the cumulative percentage : .
[0036] S43. When the cumulative percentage is not less than 99%, stop the calculation and record the corresponding number of eigenvalues at this time. , that is, the number of eigenvalues is the dimension after dimensionality reduction.
[0037] S44. Constructing transformation matrix : .
[0038] S45. Perform data dimensionality reduction to obtain the sample data set matrix after dimensionality reduction : , so far, originally dimensional data is reduced to the current dimension.
[0039] S5. Based on the sample data set after dimensionality reduction, support vector machine is used for binary classification training to obtain the separating hyperplane and binary classification model for the engine stall problem.
[0040] After the above steps, the original data samples have been reduced to dimension, and obtain a new sample data set matrix , which contains The dimensions are Sample.
[0041] According to the labels of the samples collected in S1, that is, the corresponding flameout data and non-blank data, the training set is given :
[0042] Where, is the sample after dimensionality reduction; is the sample label, , indicating the engine is off. , indicating the normal operating state of the engine.
[0043] The support vector machine method is used to classify the above samples, find the separating hyperplane, and obtain a binary classification model. The specific method includes: S51. Select soft constraint parameters , construct a quadratic programming problem under soft constraint parameters, which satisfies the expression:
[0044] Use MATLAB's built-in quadratic programming function quadprog to solve and obtain the optimal solution :
[0045] S52. Calculate the optimal weight vector and bias term: Optimal weight vector : .
[0046] choose A component of ,satisfy , then calculate the bias term : .
[0047] S53. A binary classification model is obtained, which satisfies the expression:
[0048] Where, is the optimal weight vector, is the bias term.
[0049] S6. Classify and predict the vehicle parameters collected in real time based on the binary classification model to determine whether the engine is at risk of stalling.
[0050] In a further solution, the vehicle parameters collected in real time will be Follow steps S2 to S4 to perform decentralization and dimensionality reduction, and we get ;Will Bring in The results are analyzed as follows: like =-1, the binary classification model predicts that the engine is not turned off and the vehicle continues to run as usual.
[0051] like =1, the binary classification model predicts a stall risk, triggering the vehicle control module to adjust the brake command to prevent the stall risk.
[0052] The present invention addresses the problem of engine stalling of unmanned vehicles in mines. Based on several pre-selected variables related to the engine stalling problem, the principal component analysis method is used to reduce the data dimension, thereby reducing the original sample matrix into a new sample matrix with a smaller dimension. Based on the new sample matrix with a smaller dimension, the support vector machine method is used to find the separating hyperplane and obtain a binary classification model, thereby realizing the prediction of the stalling problem. Based on the prediction result, it can be known in advance whether the current instruction will cause the engine to stall, so that effective measures can be adopted from a technical level to avoid it.
[0053] Verified by measured data from mining areas, this method has an engine stall prediction accuracy of over 92% and a false alarm rate of less than 5% at a 99% confidence level, and can effectively predict stall problems in rigid mining vehicles with automatic transmissions.
[0054] In addition, the present invention also protects an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the engine stall problem classification method are implemented.
[0055] The present invention also protects a computer-readable storage medium having a computer program stored thereon, which implements the steps of the engine stall problem classification method when the computer program is executed by a processor.
[0056] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A method for classifying engine stall problems, characterized in that: The method comprises: Collecting vehicle parameters with and without the engine turned off to form an original sample set; decentralizing the original sample set to obtain a decentralized sample matrix; wherein the vehicle parameters include vehicle speed, gear position, brake pressure, engine speed, retarder status, vehicle load, and current slope; Constructing a covariance matrix based on the centered sample matrix and calculating all eigenvalues and corresponding eigenvectors; Determine the dimension after dimensionality reduction according to the cumulative percentage of the eigenvalues, construct a transformation matrix and perform dimensionality reduction processing on the samples to generate a sample data set after dimensionality reduction; Based on the sample data set after dimensionality reduction, a support vector machine is used to perform binary classification training to obtain a separating hyperplane and a binary classification model for the engine stall problem; The vehicle parameters collected in real time are classified and predicted based on the binary classification model to determine whether the engine is at risk of stalling.
2. The engine stall problem classification method according to claim 1, characterized in that: The vehicle parameters collected when the engine is turned off and when the engine is not turned off constitute an original sample set, including: Define a vector ,in, is the vehicle speed, For the gear position, is the brake pressure, is the engine speed, is the retarder state, is the vehicle load, is the current slope; The collected The data when the engine is turned off is expressed as ; The collected The data when the engine is not turned off is expressed as ; but The original sample set composed of group data is .
3. The engine stall problem classification method according to claim 2, characterized in that: The decentralizing process of the original sample set to obtain a decentralized sample matrix specifically includes: Calculate the mean vector of the original sample set : ; Subtract the mean vector from each sample to obtain the decentralized sample : ; According to the sample , and obtain the decentralized sample matrix ,in .
4. The engine stall problem classification method according to claim 3, characterized in that: The covariance matrix is constructed based on the centered sample matrix, and all eigenvalues and corresponding eigenvectors are calculated, including: Define the covariance matrix : ; Using the Matlab function eig( ) calculated eigenvalues sorted from largest to smallest and the corresponding feature vectors .
5. The engine stall problem classification method according to claim 2, characterized in that: Determining the dimension after dimensionality reduction according to the cumulative percentage of the eigenvalues includes: Calculate the sum of the absolute values of all features : , where for The eigenvalue of the sequence number, is the number of features; Starting from the first of the mentioned eigenvalues, calculate the cumulative percentage : ; until the cumulative percentage is not less than 99%, the corresponding number of eigenvalues is the dimension after dimensionality reduction.
6. The engine stall problem classification method according to claim 1, characterized in that: The binary classification training using a support vector machine includes: Construct a quadratic programming problem under soft constraint parameters and use the quadratic programming function quadprog provided by Matlab to solve and obtain the optimal weight vector and bias term; Based on the optimal weight vector and the bias term, the separating hyperplane and the binary classification model are determined, wherein the binary classification model satisfies the expression: Where, is the optimal weight vector, is the bias term.
7. The engine stall problem classification method according to claim 6, characterized in that: Determining whether the engine has a flameout risk includes: like =-1, the binary classification model predicts that the engine is not turned off and the vehicle continues to run as usual; like =1, the binary classification model predicts a stall risk, triggering the vehicle control module to adjust the brake command to prevent the stall risk from occurring.
8. The engine stall problem classification method according to claim 1, characterized in that: The method is applied to rigid fuel mining vehicles in unmanned mine driving scenarios to predict engine stalling problems caused by vehicle braking operations.
9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the engine stall problem classification method according to any one of claims 1 to 8 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the engine stall problem classification method according to any one of claims 1 to 8 are implemented.