State evaluation method for vehicle gearbox

Abstract

The invention relates to the technical field of gear box state evaluation, and discloses a vehicle gear box state evaluation method, which comprises the following steps: S1, collecting physical and chemical performance data of oil in a gear box; s2, preprocessing the collected data through an independent component analysis (ICA) method, and screening out a data set which can be used as model input; dividing the model input data set into a test data set and a training data set; s3, improving a RBMO (Red Blue Mark Optimization) algorithm, including initializing an RBMO population through a Latin hypercube sampling method to improve the population initialization uniformity; a spiral search strategy is introduced in the RBMO foraging stage, so that the red-mouth blue-streak search efficiency is improved; in the RBMO attack stage, a weight factor is introduced to optimize the position update of the red-beaked blue-streak, and an improved red-beaked blue-streak optimization algorithm IRBMO and the like are obtained; according to the method, the defect of blindness of parameter selection in the training process is overcome, and the prediction precision of the regression prediction model is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of gearbox status assessment, and in particular to a method for assessing the status of a vehicle gearbox. Background Art

[0002] The gearbox is a core component of the vehicle power transmission system. It is usually in a complex and harsh operating environment, subject to huge impact loads, high temperatures, vibrations, and wear pressure caused by long-term operation. In high-intensity combat missions, common types of gearbox failures include gear wear, bearing fatigue failure, lubricant deterioration, and abnormal meshing of transmission components. If these faults are not discovered in time, they may lead to power interruption, vehicle shutdown, or even complete equipment failure, posing a serious threat to mission execution and combat safety. Through status assessment, the operation of the gearbox can be grasped in real time, potential failure risks can be discovered in time, and vehicle decommissioning or mission failure due to gearbox failure at critical moments can be avoided, thereby ensuring the continuous combat capability of armored vehicles in complex battlefield environments. Therefore, it is particularly important to conduct status assessment of the gearbox.

[0003] In the field of mechanical system status assessment and fault diagnosis, anomaly detection technology has been widely used in the status monitoring of industrial equipment due to its sensitivity to early faults and real-time performance. Support vector machine (SVM) is a mature machine learning algorithm with strong generalization and nonlinear processing capabilities. However, when the amount of fault data is large (such as long-term monitoring of high-frequency vibration signals), the training speed slows down significantly and it is difficult to meet the real-time requirements. Gearbox faults usually involve multi-state classification (such as normal, tooth root cracks, broken teeth, wear, etc.). SVM is a binary classifier and needs to be expanded through a "one-to-one" or "one-to-many" strategy, which may introduce category conflicts and computational redundancy.

[0004] The single-class support vector machine developed on this basis is specifically used for anomaly detection problems. OCSVM can effectively identify unknown abnormal states by using only normal data for model training, and is particularly suitable for state assessment tasks of complex mechanical systems such as gearboxes. Compared with other algorithms, the single-class support vector machine has strong adaptability to the scarcity of abnormal samples, good processing ability for nonlinear relationships, and good generalization ability, which makes the state assessment results have higher accuracy. Summary of the invention

[0005] Aiming at the above-mentioned disadvantages and deficiencies in the prior art, the present invention provides a method for evaluating the state of a vehicle gearbox, which improves the Red-Billed Blue Magpie Optimization Algorithm (RBMO). The population initialization of RBMO is carried out by the Latin Hypercube Sampling method to improve the uniformity of population initialization. A spiral search strategy is introduced in the foraging stage of RBMO to improve the search efficiency of the red-billed blue magpie. A weight factor is introduced in the attack stage of RBMO to optimize the position update of the red-billed blue magpie, and the key parameters of the One-Class Support Vector Machine (OCSVM) are optimized to construct an IRBMO-OCSVM fault diagnosis model, which makes up for the defect of blindness in parameter selection during the training process. The method includes the following steps: S1. Collect the physical and chemical property data of the oil in the gearbox; S2. Preprocess the collected data through the Independent Component Analysis (ICA) method to screen out the data set that can be used as the model input. The model input data set is divided into a test data set and a training data set; S3. Improve the Red-Billed Blue Magpie Optimization Algorithm (RBMO), including initializing the RBMO population by the Latin Hypercube Sampling method to improve the uniformity of population initialization; introducing a spiral search strategy in the foraging stage of RBMO to improve the search efficiency of the red-billed blue magpie; introducing a weight factor in the attack stage of RBMO to optimize the position update of the red-billed blue magpie, and obtaining the improved Red-Billed Blue Magpie Optimization Algorithm (IRBMO); S4. Use the improved Red-Billed Blue Magpie Optimization Algorithm (IRBMO) to optimize the key parameters of the One-Class Support Vector Machine (OCSVM), construct an IRBMO-OCSVM fault diagnosis model, and train the fault diagnosis model through the training data set in S2; S5. Input the test data set in S2 into the trained IRBMO-OCSVM model to diagnose the faults of the gearbox and output the prediction results.

[0006] In step S3, the RBMO population is initialized by the Latin Hypercube Sampling method, and the formula is: ; ; In the formula: a i and b i are the two ends of an interval [a i , b i ; is the length of the i-th sub-interval when the interval [a i , b i is divided into n sub-intervals; k is a random number from 1 to n; x ij is a point selected from the k-th sub-interval; rand(0,1) is a random number uniformly distributed between [0,1]; Introduce a spiral search strategy during the foraging stage of RBMO to optimize the position update of red-billed blue magpies. The formula is as follows: 0 ≤ Rand 1 < 0.5; 0.5 ≤ Rand 1 ≤ 1; In the formula, t represents the current iteration number; X i (t + 1) represents the position of the i-th individual at the (t + 1)-th iteration. p is a random number between [2, 5] representing the number of red-billed blue magpies in a small group, q is a random number between [10, n] representing the number of red-billed blue magpies when forming a group, X m represents the randomly selected m-th individual, X i represents the i-th individual, X rs (t) represents the randomly selected individual in the current iteration; Rand 1 is a random number; is an exponential function used to adjust the expansion or contraction degree of the spiral. c and l respectively control the tightness and periodic change of the spiral; Introduce a weight factor during the attack stage of RBMO to optimize the position update of red-billed blue magpies. The formula is as follows: 0 ≤ Rand n < 0.5; 0.5 ≤ Rand n ≤ 1; ; ; In the formula, X food(t) represents the position of the food; Rand n represents a random number used to generate a standard normal distribution; T is the maximum number of iterations; CF is a dynamic variable that changes with the number of iterations; is the weight factor; Finally, judge the global optimal value. The formula is as follows: ; Among them and respectively represent the fitness values before and after the position update of the i-th red-billed blue magpie.

[0007] The improved red-billed blue magpie optimization algorithm IRBMO in step S4 optimizes the key parameters of the one-class support vector machine OCSVM. The optimized parameters include the kernel parameter, regularization parameter, and relaxation variable.

[0008] Compared with the prior art, the present invention has the following beneficial technical effects and advantages: The independent component analysis method is used to preprocess the data, which can extract independent key signal features, separate the non-linear characteristics between the physical and chemical information of the oil fluid, remove the noise in the data, and improve the data quality; The improved red-billed blue magpie optimization algorithm IRBMO is used to optimize the key parameters of the one-class support vector machine OCSVM, and an IRBMO–OCSVM fault diagnosis model is constructed, which makes up for the defect of blindness in parameter selection during the training process and improves the prediction accuracy of the regression prediction model; the population initialization of RBMO is carried out by the Latin hypercube sampling method to improve the uniformity of population initialization; a spiral search strategy is introduced in the foraging stage of RBMO to improve the search efficiency of the red-billed blue magpie; a weight factor is introduced in the attack stage of RBMO to optimize the position update of the red-billed blue magpie and avoid falling into local optima. Specific implementation manner

[0009] The present invention will be described in detail below: The present invention provides a method for evaluating the state of a vehicle gearbox, including the following steps: S1. Collect the physical and chemical property data of the oil fluid in the gearbox.

[0010] S2. Preprocess the collected data through the independent component analysis method ICA, separate the non-linear characteristics between the physical and chemical information of the oil fluid, remove the noise in the data, improve the data quality, and screen out the data set that can be used as the model input; divide the model input data set into a test data set and a training data set.

[0011] Specifically, it includes the following steps: S201. Centralize the data and whiten the data: This step aims to simplify the separation process of independent components. By centralizing and whitening the preprocessed data, it has zero mean and unit variance, and the components are uncorrelated with each other.

[0012] Centralize the data: The mean of each variable (data dimension) is subtracted so that the mean of the processed data is zero. The formula is: (1), where is the data matrix after centralization processing; μ is the mean of the original data X.

[0013] Whiten the data: The data is transformed into a new space so that its components are independent of each other and the variance is 1.

[0014] The formula is: (2), where X white is the data matrix after whitening processing; E and D are the eigenvector and eigenvalue matrix of the data covariance matrix respectively.

[0015] S202. Optimization objective and application of ICA algorithm: Recover statistically independent source signals from the whitened data. This is achieved by optimizing an objective function that quantifies the statistical independence of the components.

[0016] The most common method for the optimization objective is to maximize non-Gaussianity, because according to the central limit theorem, the stronger the non-Gaussianity, the higher the independence; the commonly used non-Gaussianity measure is the fourth-order cumulant (kurtosis) or negentropy.

[0017] Select a non-linear function to approximate the non-Gaussianity of the independent components; Iteratively update the weight matrix W: , where W is the weight matrix; g is the non-linear function; g′ is the derivative of g; {} represents the expectation; W (next) is the demixing matrix updated during the ICA iteration process.

[0018] S203. Normalization and rotation: Normalize the weight matrix W and possibly perform an orthogonal rotation to further improve the independence of the sources.

[0019] Normalization formula: , ensures the stability and convergence results of the weight matrix; Finally, the output of the ICA model: , S represents the independent components in the original data, where W T is the demixing matrix learned from the data.

[0020] Through the above steps, ICA can extract meaningful signals from complex multi-variable data, and these signals are as statistically independent as possible.

[0021] S3. Improve the Red-billed Blue Magpie Optimization Algorithm (RBMO), including initializing the RBMO population by Latin hypercube sampling to improve the uniformity of population initialization; introducing a spiral search strategy in the foraging stage of RBMO to improve the search efficiency of the red-billed blue magpie; introducing a weight factor in the attack stage of RBMO to optimize the position update of the red-billed blue magpie, and obtaining the improved Red-billed Blue Magpie Optimization Algorithm (IRBMO).

[0022] Specifically, in step S3, the RBMO population is initialized by the Latin hypercube sampling method. Introducing the Latin hypercube sampling method to initialize the population, since the Latin hypercube sampling can generate sample points in a multi-dimensional space, it is an improved stratified sampling technique that can ensure uniform and representative coverage of the entire range of each parameter. It helps to improve the global search ability of the red-billed blue magpie algorithm. The formula is as follows: (6); (7), where: a i and b i are the two ends of an interval [a i , b i ; is the length of the i-th sub-interval when the interval [a i , b i is divided into n sub-intervals; k is a random number from 1 to n; x ij is a point selected from the k-th sub-interval; rand(0,1) is a random number uniformly distributed between [0,1].

[0023] During the foraging stage of RBMO, that is, when looking for food, red-billed blue magpies usually act in small groups (2 to 5 individuals) or in flocks (more than 10 individuals) to improve the search efficiency.

[0024] Set a random number Rand 1 , when 0 ≤ Rand 1 < 0.5, the red-billed blue magpie population forms multiple small groups for search, and the formula is: , 0 ≤ Rand1 < 0.5; (8), When 0 ≤ Rand 1 < 0.5, the red-billed blue magpie population searches in multiple flocks, and the formula is: , 0.5 ≤ Rand1 ≤ 1; (9), where, t represents the current iteration number; X i (t + 1) represents the position of the i-th individual at the (t + 1)-th iteration, p is a random number between [2, 5] representing the number of red-billed blue magpies in a small group, q is a random number between [10, n] representing the number of red-billed blue magpies in a flock, X m represents the randomly selected m-th individual, X i represents the i-th individual, and X rs (t) represents the randomly selected individual in the current iteration.

[0025] The spiral search strategy can effectively expand the search space, especially in complex or multimodal optimization problems, helping the algorithm explore multiple potential good regions, thereby increasing the likelihood of finding the global optimal solution. Introducing the spiral search strategy in the foraging stage of RBMO optimizes the position update of red-billed blue magpies. The update formulas for red-billed blue magpies to search in small groups and in flocks are as follows: 0 ≤ Rand 1 < 0.5; (10), 0.5 ≤ Rand 1 ≤ 1; (11), In the formula, is the exponential function used to adjust the expansion or contraction degree of the spiral, and c and l respectively control the tightness and periodic change of the spiral.

[0026] In the attack stage of RBMO, a random number Rand n is set. When 0 ≤ Rand n < 0.5, red-billed blue magpies act in small groups, and their main targets are usually small prey or plants. The corresponding mathematical model is as follows: , 0 ≤ Rand n < 0.5; (12).

[0027] When 0.5 < Rand n ≤ 1, red-billed blue magpies act in flocks, and they can jointly target large prey, such as large insects or small vertebrates. The corresponding mathematical model is as follows: 0.5 ≤ Rand n ≤ 1; (13), ; (14), In the formula, X food(t) represents the position of the food; Rand n represents the random number used to generate the standard normal distribution; T is the maximum number of iterations; CF is a dynamic variable that changes with the number of iterations.

[0028] At this time, a weight factor is introduced to optimize the position update of red-billed blue magpies, enabling each individual position to communicate with other individuals in the local search space when moving towards the global optimal position, making the local search ability of the algorithm stronger. In this embodiment, a weight factor with a sinusoidal change is selected to control the influence of the prey target on the position update of red-billed blue magpies. The formula is: 0 ≤ Rand n < 0.5; (15), 0.5 < Rand n ≤ 1; (16), ; (17), Among them, is the weight factor.

[0029] Finally, judge the global optimal value: In addition to foraging and attacking food, the red-billed blue magpie also stores the excess food. This process preserves the information of the solution, facilitating an individual to find the global optimal value. The formula is as follows: ; (18), where and represent the fitness values of the $i$-th red-billed blue magpie before and after position update, respectively.

[0030] S4. Use the improved red-billed blue magpie optimization algorithm IRBMO to optimize the key parameters of the one-class support vector machine OCSVM, including the kernel parameter, regularization parameter, and slack variable, and construct the IRBMO–OCSVM fault diagnosis model. Train the fault diagnosis model using the training data set in S2.

[0031] The specific process of parameter optimization is as follows: (1) Initialize the parameters: kernel parameter, regularization parameter, slack variable.

[0032] (2) Initialize the population according to formula (6), calculate the fitness values of the population individuals, and determine the optimal position.

[0033] (3) Conduct foraging behavior. If $0\leq Rand$ 1 $< 0.5$, update the position information of the individual according to formula (10); if $0.5\leq Rand$ 1 $\leq 1$, update the position information of the individual according to formula (11).

[0034] (4) Conduct attack behavior. If $0\leq Rand_n < 0.5$ , update the position information of the individual according to formula (15) and conduct small-group attack behavior; if $0.5\leq Rand_n < 1$ , update the individual information according to formula (16) and conduct attack behavior in a group manner.

[0035] (5) Compare the fitness values before and after update. If then update the optimal individual position; otherwise, do not update the individual position.

[0036] (6) Judge whether the algorithm meets the end condition, that is, whether the maximum number of iterations is reached. If the maximum number of iterations is reached, output the optimal parameters; if the maximum number of iterations is not reached, repeat steps (3) - (5).

[0037] (7) Input the output optimal parameter combination into OCSVM.

[0038] S5. Input the test data set in S2 into the trained IRBMO–OCSVM model to conduct fault diagnosis on the gearbox and output the prediction results.

[0039] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Any modifications, alterations, substitutions, and variations made by those of ordinary skill in the art to the above embodiments fall within the scope of the present invention.

[0040] The above description is only for the preferred embodiments of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for evaluating the condition of a vehicle gearbox, characterized in that: The following steps are involved: S1. Collect the physical and chemical properties data of the oil in the gearbox; S2. Preprocess the collected data through independent component analysis (ICA) to select the data set that can be used as model input; divide the model input data set into a test data set and a training data set; S3. Improvements to the red-billed blue magpie optimization algorithm RBMO, including initializing the RBMO population through the Latin hypercube sampling method to improve the uniformity of population initialization; introducing a spiral search strategy in the RBMO foraging phase to improve the search efficiency of the red-billed blue magpie; introducing a weight factor in the RBMO attack phase to optimize the position update of the red-billed blue magpie, and obtaining an improved red-billed blue magpie optimization algorithm IRBMO; S4, using the improved red-billed blue magpie optimization algorithm IRBMO to optimize the key parameters of the single-class support vector machine OCSVM, constructing the IRBMO-OCSVM fault diagnosis model, and training the fault diagnosis model with the training data set in S2; S5. Input the test data set in S2 into the trained IRBMO–OCSVM model, perform fault diagnosis on the gearbox, and output the prediction results.

2. A vehicle gearbox condition assessment method according to claim 1, characterized in that: In step S3, the RBMO population is initialized by the Latin hypercube sampling method, and the formula is: ； ； Where: a i and b i For an interval [a i , b i ] at both ends; is to divide the interval [a i , b i ] is divided into the length of the i-th subinterval among n subintervals; k is a random number between 1 and n; x ij is a point selected from the kth subinterval; rand(0,1) is a random number uniformly distributed between [0,1]; In the RBMO foraging stage, the spiral search strategy is introduced to optimize the position update of the red-billed blue magpie: 0≤Rand1<0.5; 0.5≤Rand1≤1; Where, t is the current iteration number; X i (t+1) represents the position of the i-th individual in the t+1 iteration, p is a random number between [2,5], representing the number of red-billed blue magpies in a small group, q is a random number between [10,n], representing the number of red-billed blue magpies in a group, X m represents the randomly selected mth individual, X i represents the i-th individual, X rs (t) represents the randomly selected individual in the current iteration; Rand1 is a random number; It is an exponential function used to adjust the expansion or contraction of the spiral, and c and l control the tightness and periodicity of the spiral respectively; In the RBMO attack phase, the weight factor is introduced to optimize the position update of the red-billed blue magpie: , 0≤Rand n<0.5; , 0.5≤Rand n≤1; ； ； Where, X food(t) Indicates the location of the food; Rand n represents the random number used to generate the standard normal distribution; T is the maximum number of iterations; CF is a dynamic variable that changes with the number of iterations; is the weight factor; Finally, the global optimal value is determined as follows: ； in and They respectively represent the fitness values ​​of the i-th red-billed blue magpie before and after the position is updated.

3. The method for evaluating the condition of a vehicle gearbox according to claim 1, characterized in that: The improved red-billed blue magpie optimization algorithm IRBMO in step S4 optimizes key parameters of the single-class support vector machine OCSVM, and the optimized parameters include kernel parameters, regularization parameters, and slack variables.

Images