An AMT clutch wear state online diagnosis method and system

Abstract

The application provides an AMT clutch wear state online diagnosis method and system, and belongs to the technical field of clutch wear diagnosis. The method comprises the following steps: triggering wear state diagnosis in the sliding film stage, the combination stage and the full combination stage of the clutch, and acquiring vibration acceleration original signals collected continuously in time at the AMT clutch pressure plate; extracting time domain statistical features related to transient impact components and energy levels in the vibration acceleration original signals, and constructing a multi-dimensional feature vector; and inputting the multi-dimensional feature vector into a pre-trained support vector machine to obtain a diagnosis result of the clutch wear state. The method analyzes the time domain statistical features of the clutch vibration signals, uses the trained vector machine model to realize online diagnosis and classification of the clutch wear state, and achieves the purpose of real-time and accurate AMT clutch fault diagnosis.

Description

Technical Field

[0001] This invention relates to the field of clutch wear diagnosis technology, specifically to an online diagnostic method and system for AMT clutch wear condition. Background Technology

[0002] In the commercial vehicle sector, automatic transmission models are gaining an increasing market share in China, while AMT automatic transmissions are favored by consumers for their high transmission efficiency and strong fuel economy.

[0003] As AMT users' mileage increases, the clutch of the AMT will inevitably experience severe wear. Unlike manual transmission vehicles, where drivers can directly feel the clutch's condition when shifting gears, AMT users often neglect to perform periodic clutch checks. This can lead to vehicle breakdowns when the clutch malfunctions, causing inconvenience to customers. Furthermore, clutch inspection during after-sales service requires disassembly and disassembly, a cumbersome process that results in prolonged vehicle downtime and significantly increases maintenance costs. Summary of the Invention

[0004] This invention provides an online diagnostic method for AMT clutch wear conditions. This method analyzes the time-domain statistical characteristics of clutch vibration signals and utilizes a trained vector machine model to achieve online diagnosis and classification of clutch wear conditions, thus achieving real-time and accurate AMT clutch fault diagnosis. The technical solution provided by this invention is as follows:

[0005] This invention provides an online diagnostic method for the wear condition of an AMT clutch, comprising the following steps: Wear condition diagnosis is triggered during the slippage stage, engagement stage, and full engagement stage of the clutch, and the raw vibration acceleration signal collected continuously at the AMT clutch pressure plate is obtained. Extract time-domain statistical features related to transient impact components and energy levels from the original vibration acceleration signal, and construct a multi-dimensional feature vector; The multidimensional feature vectors are input into a pre-trained support vector machine to obtain diagnostic results of clutch wear status.

[0006] Preferably, the time-domain statistical characteristics related to the transient impact component and energy level include: Root mean square (RMS) is used to characterize the average energy of a signal. ; Peak value, used to characterize the maximum impulse amplitude in a signal: ; Kurtosis is used to characterize the degree of sensitivity to shock components. ; Peak factor, used to identify impact failures: ; Pulse factor, used to characterize sensitivity to transient pulses: ; Margin factor, an effective indicator for measuring the severity of an impact: ; Skewness, used to describe the asymmetry of signal distribution: ; Variance is used to characterize the intensity of signal fluctuations. ; Therefore, the multidimensional feature vector is: ; In the formula, N The number of data points for each sample; For the first i The vibration signal amplitude at each sample point; The mean of the sample. ; The standard deviation of the sample. .

[0007] Preferably, the training of the support vector machine includes the following steps: Several AMT automatic transmissions equipped with clutches were assembled on the test bench. The clutches were divided into several clutches with slight wear, moderate wear, and severe wear. Vibration acceleration sensors are installed on each clutch pressure plate to collect raw vibration acceleration signals over continuous time. A fixed length is extracted from the raw vibration acceleration signal acquired over a continuous time period. N The samples are used as analysis samples, and each sample is labeled with a wear category, including slight wear, moderate wear, and severe wear; training sets and test sets are constructed separately. The original vibration acceleration signals in the training set are subjected to time-domain statistical feature extraction to obtain multi-dimensional feature vectors. The feature vectors are then input into the support vector machine model for iterative training until the set error is met. The accuracy of the trained support vector machine model is tested using a test set. If the accuracy meets the set target, the trained support vector machine model is obtained.

[0008] The present invention also provides an online diagnostic system for the wear condition of an AMT clutch, the system comprising: The signal acquisition module includes two vibration acceleration sensors, which are symmetrically arranged on each clutch pressure plate to collect raw vibration acceleration signals in real time during the slip phase, engagement phase and full engagement phase of the clutch. The signal processing module is used to extract time-domain statistical features related to transient impact components and energy levels from the original vibration acceleration signal and construct a multi-dimensional feature vector. The clutch condition diagnosis module is used to obtain the diagnosis results of clutch wear condition based on multidimensional feature vectors and pre-trained support vector machines. The warning output module is used to output diagnostic results of clutch wear status and send them to the vehicle dashboard and remote data platform via CAN messages.

[0009] The beneficial effects of this invention are: This invention proposes an online diagnostic method for the wear condition of an AMT clutch. This method deeply analyzes the vibration acceleration information of the clutch pressure plate during the slip phase, engagement phase, and full engagement phase. It extracts the transient impact components and energy level-related time-domain statistical features of the clutch under operating conditions to reflect the wear condition of the clutch under the current operating conditions. Furthermore, it employs a support vector machine to quickly and accurately diagnose the wear condition of the clutch, which can significantly improve the accuracy of clutch condition identification.

[0010] This invention allows for online assessment of the clutch's health and wear status without disassembling the transmission, transforming AMT after-sales service from a reactive response to proactive planning and significantly reducing AMT after-sales repair rates.

[0011] The feature extraction method proposed in this invention uses time-domain statistical feature calculation, which can not only accurately reflect the current wear status of the clutch, but also has relatively simple time-domain statistical feature calculation, low requirements for controller computing power, and can meet the real-time diagnostic needs in the vehicle environment. Attached Figure Description

[0012] Figure 1 This is a flowchart of the AMT clutch wear diagnosis method according to an embodiment of the present invention; Figure 2 This is an architecture diagram of the AMT clutch wear condition diagnosis system according to an embodiment of the present invention; Figure 3 This is a flowchart of the support vector machine sample training process according to an embodiment of the present invention; Figure 4 This is a confusion matrix diagram for AMT clutch wear condition diagnosis and identification according to an embodiment of the present invention; Figure 5 This is a graph showing the prediction results of test samples in an embodiment of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0014] Example 1 This invention proposes an online diagnostic method for the wear condition of an AMT clutch, the flowchart of which is shown below. Figure 1 As shown, the specific steps are as follows: S1: Triggers wear condition diagnosis during the slip phase, engagement phase, and full engagement phase of the clutch.

[0015] S2: Acquire the raw vibration acceleration signal continuously collected at the AMT clutch pressure plate.

[0016] S3: Extract time-domain statistical features related to transient impact components and energy levels from the original vibration acceleration signal and construct a multi-dimensional feature vector.

[0017] Specifically, the time-domain statistical characteristics related to the transient impact components and energy levels include: Root mean square (RMS) is used to characterize the average energy of a signal. ; Peak value, used to characterize the maximum impulse amplitude in a signal: ; Kurtosis is used to characterize the degree of sensitivity to shock components. ; Peak factor, used to identify impact failures: ; Pulse factor, used to characterize sensitivity to transient pulses: ; Margin factor, an effective indicator for measuring the severity of an impact: ; Skewness, used to describe the asymmetry of signal distribution: ; Variance is used to characterize the intensity of signal fluctuations. ; Therefore, the multidimensional feature vector is: ; In the formula, N The number of data points for each sample; For the first i The vibration signal amplitude at each sample point; The mean of the sample. ; The standard deviation of the sample. .

[0018] S4: Input the multidimensional feature vector into the pre-trained support vector machine to obtain the diagnostic results of the clutch wear state.

[0019] The training of a support vector machine includes the following steps: Several AMT automatic transmissions equipped with clutches were assembled on the test bench. The clutches were divided into several clutches with slight wear, moderate wear, and severe wear. Vibration acceleration sensors are installed on each clutch pressure plate to collect raw vibration acceleration signals over continuous time. A fixed length is extracted from the raw vibration acceleration signal acquired over a continuous time period. N ( N A sample of 1024 data points is used as an analysis sample. Each sample is labeled with a wear category, including slight wear, moderate wear, and severe wear. Training and test sets are constructed separately. Time-domain statistical features are extracted from the original vibration acceleration signals of the training set to obtain multi-dimensional feature vectors. The feature vectors are input into the support vector machine model for iterative training until the set error is met. The accuracy of the trained support vector machine model is tested using the test set. If the set accuracy is met, the trained support vector machine model is obtained.

[0020] In this embodiment, as Figure 1 As shown, during real-time diagnostics, symmetrical vibration acceleration sensors are arranged at the AMT clutch pressure plate to collect vibration signals and obtain the raw signals. x ( t The obtained raw signal x ( t Perform time-domain statistical feature extraction to obtain feature vectors. F , to feature vector F The data is input into a pre-trained vector machine model for clutch state recognition and diagnosis.

[0021] Specifically, when constructing the training and test sets, several clutches of three types were installed on the AMT test bench: State 1, clutches with slight wear after less than 50,000 kilometers of operation; State 2, clutches with moderate wear after 50,000-150,000 kilometers of operation; and State 3, clutches with severe wear and obvious burning after more than 150,000 kilometers of operation. The test bench simulated the clutch slippage process during AMT vehicle start-up, the clutch engagement process during gear shifting, and the clutch full engagement process during normal AMT vehicle operation. Vibration signals from clutches at different wear levels and in these three stages were collected. Each vibration signal was collected as a sample of 1024 data points, with 200 samples collected for each wear level and each stage, totaling 3 x 3 x 200 = 1800 samples. These 1800 samples were divided into two groups at a 5:1 ratio: the first group (1500 samples) was the training set, and the second group (300 samples) was the test set. Temporal statistical feature extraction is performed on the sample signals of the training and test sets to obtain feature vectors. F For eigenvectors F Normalization is performed to eliminate some excessively large vector values, so that all features are on the same order of magnitude.

[0022] like Figure 3 As shown, the training set is input into the vector machine model for training. The basic principle of the vector machine is based on margin maximization. For a linearly separable dataset, the goal of the vector machine is to find a hyperplane that separates data points of different classes and maximizes the distance between the hyperplane and the nearest data point. During training, the vector machine receives the feature matrix (1500x8) and the corresponding label vector (1500x1) of the training set, and adjusts the parameters through the radial basis function kernel and the penalty parameter C. Training is completed when the training error rate is less than 0.05.

[0023] like Figure 4 As shown, after training, the 1500 samples in the training set were used to identify 500 cases of slight wear, 500 cases of moderate wear, and 500 cases of severe wear in the actual results. The vast majority of these cases were successfully identified by the vector machine, with an accuracy rate of (490+485+487) / 1500=97.5%.

[0024] The input test set is used to diagnose and identify the clutch wear state of the trained vector machine model. Slight wear is represented by 1, moderate wear by 2, and severe wear by 3. Figure 5 The prediction results of the test set shown have an accuracy of approximately 98% when compared with known actual results.

[0025] It should be noted that during driving, the AMT will actively perform an online diagnostic identification of the clutch wear status every 10,000 kilometers. It will collect signals from the clutch slip stage, engagement stage, and full engagement stage for online analysis, and send the identification results (clutch wear degree: slight wear / moderate wear / severe wear) to the vehicle instrument panel and remote data platform via CAN bus message.

[0026] The above is an embodiment of the online diagnostic method for AMT clutch wear condition provided in this example. Based on the same idea, this example also provides a corresponding online diagnostic system for AMT clutch wear condition, the system architecture of which is as follows: Figure 2 As shown, it specifically includes: The signal acquisition module includes two vibration acceleration sensors, which are symmetrically arranged on each clutch pressure plate to collect raw vibration acceleration signals in real time during the slip phase, engagement phase and full engagement phase of the clutch.

[0027] The signal processing module is used to extract time-domain statistical features related to transient impact components and energy levels from the original vibration acceleration signal and construct a multi-dimensional feature vector.

[0028] The clutch condition diagnosis module is used to obtain diagnostic results of clutch wear condition based on multidimensional feature vectors and pre-trained support vector machines.

[0029] The warning output module is used to output diagnostic results of clutch wear status and send them to the vehicle dashboard and remote data platform via CAN messages.

[0030] The modules in the aforementioned online diagnostic system for AMT clutch wear can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0031] This embodiment also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 Provides an online diagnostic method for AMT clutch wear condition.

[0032] Those skilled in the art will understand that implementing all or part of the processes in the methods of the above embodiments can be accomplished by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0033] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An online diagnostic method for the wear condition of an AMT clutch, characterized in that, Includes the following steps: Wear condition diagnosis is triggered during the slippage stage, engagement stage, and full engagement stage of the clutch, and the raw vibration acceleration signal collected continuously at the AMT clutch pressure plate is obtained. Extract time-domain statistical features related to transient impact components and energy levels from the original vibration acceleration signal, and construct a multi-dimensional feature vector; The multidimensional feature vectors are input into a pre-trained support vector machine to obtain diagnostic results of clutch wear status.

2. The online diagnostic method for AMT clutch wear condition according to claim 1, characterized in that, The time-domain statistical characteristics related to the transient impact components and energy levels include: Root mean square (RMS) is used to characterize the average energy of a signal. ； Peak value, used to characterize the maximum impulse amplitude in a signal: ； Kurtosis is used to characterize the degree of sensitivity to shock components. ； Peak factor, used to identify impact failures: ； Pulse factor, used to characterize sensitivity to transient pulses: ； Margin factor, an effective indicator for measuring the severity of an impact: ； Skewness, used to describe the asymmetry of signal distribution: ； Variance is used to characterize the intensity of signal fluctuations. ； Therefore, the multidimensional feature vector is: ; In the formula, N The number of data points for each sample; For the first i The vibration signal amplitude at each sample point; The mean of the sample. ; The standard deviation of the sample. .

3. The online diagnostic method for AMT clutch wear condition according to claim 1, characterized in that, The training of the support vector machine includes the following steps: Several AMT automatic transmissions equipped with clutches were assembled on the test bench. The clutches were divided into several clutches with slight wear, moderate wear, and severe wear. Vibration acceleration sensors are installed on each clutch pressure plate to collect raw vibration acceleration signals over continuous time. A fixed length is extracted from the raw vibration acceleration signal acquired over a continuous time period. N The samples are used as analysis samples, and each sample is labeled with a wear category, including slight wear, moderate wear, and severe wear; training sets and test sets are constructed separately. The original vibration acceleration signals in the training set are subjected to time-domain statistical feature extraction to obtain multi-dimensional feature vectors. The feature vectors are then input into the support vector machine model for iterative training until the set error is met. The accuracy of the trained support vector machine model is tested using a test set. If the accuracy meets the set target, the trained support vector machine model is obtained.

4. An online diagnostic system for AMT clutch wear condition, characterized in that, The system includes: The signal acquisition module includes two vibration acceleration sensors, which are symmetrically arranged on each clutch pressure plate to collect raw vibration acceleration signals in real time during the slip phase, engagement phase and full engagement phase of the clutch. The signal processing module is used to extract time-domain statistical features related to transient impact components and energy levels from the original vibration acceleration signal and construct a multi-dimensional feature vector. The clutch condition diagnosis module is used to obtain the diagnosis results of clutch wear condition based on multidimensional feature vectors and pre-trained support vector machines. The warning output module is used to output diagnostic results of clutch wear status and send them to the vehicle dashboard and remote data platform via CAN messages.

Images