Engine fault dynamic diagnosis system and method for automobile maintenance

Abstract

The invention discloses an automobile maintenance engine fault dynamic diagnosis system and method. The system comprises a data acquisition module, a data preprocessing module, a dynamic monitoring and recording module, a sound signal extraction algorithm module, a wavelet transform algorithm module, a neural network module and a fault diagnosis and prediction module. The method comprises the following steps: step 1, data acquisition; step 2, data preprocessing; step 3, real-time monitoring and recording; step 4, sound signal extraction and analysis; step 5, fault feature extraction; step 6, constructing and training a neural network; step 7, fault diagnosis and prediction; the dynamic monitoring function is designed, the vibration frequency, surface abnormity and working condition parameters of the automobile engine can be comprehensively monitored, real-time alarming is carried out on abnormal conditions, potential risks existing in the engine can be found in time, powerful support is provided for fault prevention and processing, and the working efficiency is improved. And a fast cyclic modulation spectrum peak graph algorithm is designed through a multi-stage decomposition method.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile maintenance, and in particular to a dynamic diagnosis system and method for automobile maintenance engine failure. Background Art

[0002] The engine is the heart of the car. Its performance directly affects the car's power, economy, and reliability. Therefore, regular engine maintenance is crucial. This not only extends the engine's service life, but also improves the car's driving performance and safety.

[0003] Existing automotive engine fault diagnosis methods suffer from the following drawbacks: First, they lack a dynamic engine fault diagnosis mechanism, making it impossible to continuously monitor various engine operating parameters in real time. This makes it difficult to promptly detect and warn of potential engine fault risks. The lack of dynamic diagnosis means that many initial, minor but critical fault signs may be overlooked, increasing the risk of fault deterioration and repair costs. Second, the traditional spectral kurtosis method exhibits significant limitations when applied to detecting periodic pulse signals. When dealing with complex signal environments, this method has difficulty effectively distinguishing between single pulse signals and random noise, which often leads to false positives or missed negatives. In addition, the spectral kurtosis method is extremely sensitive to the signal-to-noise ratio (SNR). When the SNR is not ideal, its diagnostic effectiveness is greatly reduced, limiting its widespread application in various practical scenarios. Third, traditional engine fault diagnosis methods overly rely on the technician's personal experience and intuition. This reliance not only makes the diagnostic results susceptible to subjective factors, but also often leads to low diagnostic efficiency and difficulty in ensuring accuracy. The lack of a standardized and intelligent diagnostic process makes the fault diagnosis process lack consistency and repeatability. Summary of the Invention

[0004] The purpose of the present invention is to provide a system and method for dynamic diagnosis of automobile engine failures to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a dynamic diagnosis system for automobile maintenance engine faults, comprising a data acquisition module, the data acquisition module being connected to a data preprocessing module, the data preprocessing module being connected to a dynamic monitoring and recording module and an acoustic signal extraction algorithm module, the acoustic signal extraction algorithm module being connected to a wavelet transform algorithm module, the wavelet transform algorithm module being connected to a neural network module, and the neural network module being connected to a fault diagnosis and prediction module.

[0006] As a further technical solution of the present invention, the data acquisition module includes a vibration frequency acquisition module, a surface anomaly monitoring module, an operating parameter acquisition module and an acoustic signal acquisition module, and the data preprocessing module includes a signal filtering module, a data calibration module and a data normalization module.

[0007] As a further technical solution of the present invention, the dynamic monitoring and recording module includes a real-time data monitoring module, a data recording and storage module, and an abnormal alarm module, and the acoustic signal extraction algorithm module includes a multi-level decomposition module, a cyclic modulation spectrum peak calculation module, a fast search strategy module, and a periodic pulse signal extraction module.

[0008] As a further technical solution of the present invention, the wavelet transform algorithm module includes a multi-scale decomposition module and a fault feature extraction module, and the neural network module includes a neural network construction module, an input layer node configuration module, an output layer node configuration module, a data set module, a back propagation algorithm training module and a network weight adjustment module.

[0009] As a further technical solution of the present invention, the fault diagnosis and prediction module includes a fault identification module, a fault type judgment module, a fault assessment module and a fault prediction module.

[0010] As a further technical solution of the present invention, the dynamic monitoring and recording module is data-connected with a user interface and interaction module, which includes a data visualization module, a log recording module, a report generation module and a system parameter setting module, and the user interface and interaction module establishes a data connection with the fault diagnosis and prediction module.

[0011] A method for dynamic diagnosis of automobile engine failures in maintenance, comprising the following steps: data collection; data preprocessing; real-time monitoring and recording; acoustic signal extraction and analysis; fault feature extraction; neural network construction and training; and fault diagnosis and prediction.

[0012] In the above step 1, the vibration frequency, surface anomaly, operating parameters and acoustic signals of the engine are collected through the data acquisition module;

[0013] In the above step 2, the collected raw data is processed by a data preprocessing module;

[0014] In step 3 above, the dynamic monitoring and recording module is used to monitor and record the pre-processed data in real time, continuously track the changes in various engine parameters, set alarm thresholds, and issue an alarm in a timely manner when abnormal data is detected;

[0015] In the above step 4, the acoustic signal extraction algorithm module accurately extracts the engine's periodic pulse signal from the complex acoustic environment, providing key information for fault diagnosis;

[0016] In the above step 5, the wavelet transform algorithm module is used to perform multi-scale decomposition on the pre-processed data to extract the fault characteristics;

[0017] In the above step 6, a multilayer perceptron neural network is constructed according to the number of extracted fault features;

[0018] In the above step seven, the trained neural network is used to perform fault diagnosis and prediction on the real-time monitored data.

[0019] As a further technical solution of the present invention, in step one, the engine's vibration frequency, surface anomalies, operating parameters and acoustic signals are collected through a data acquisition module. The vibration frequency acquisition module uses a vibration sensor to accurately collect the engine's vibration frequency under different operating conditions, providing basic data for analyzing the engine's operating status. The surface anomaly monitoring module uses visual detection and infrared thermal imaging technology to monitor whether the engine surface has abnormal wear, cracks or temperature abnormalities. The operating parameter acquisition module collects the engine's speed, oil pressure, water temperature, oil temperature and other key operating parameters to fully understand the engine's working status. The acoustic signal acquisition module uses an acoustic sensor to collect the sound signal when the engine is running, providing original data for subsequent acoustic signal analysis.

[0020] As a further technical solution of the present invention, in step four, the multi-level decomposition module in the acoustic signal extraction algorithm module decomposes the acoustic signal into signals of different frequency bands through a multi-level decomposition method, and then calculates the cyclic modulation spectrum peak in each frequency band through the cyclic modulation spectrum peak calculation module, and locates the center frequency of the periodic pulse signal by finding the maximum CMSP. In the process, a fast search strategy is adopted by the fast search strategy module to reduce the computational complexity, improve the algorithm efficiency, and ensure real-time performance. Finally, the periodic pulse signal extraction module accurately extracts the engine's periodic pulse signal from a complex acoustic environment based on the located center frequency, providing key information for fault diagnosis.

[0021] As a further technical solution of the present invention, in step six, the neural network construction module in the neural network module constructs a multilayer perceptron neural network according to the number of extracted fault features, including an input layer, a hidden layer and an output layer. The neural network nodes are configured through the input layer node configuration module and the output layer node configuration module. The number of input layer nodes is determined according to the number of extracted fault features, and the output layer nodes correspond to different fault types. The data set module collects a large amount of known fault data as training samples to ensure the diversity and representativeness of the data. The back propagation algorithm training module uses the back propagation algorithm to train the network and adjust the network weights so that the network can accurately identify the fault type. The network weight adjustment module fine-tunes the network weights according to the training results and network performance to improve the diagnostic accuracy and robustness of the network.

[0022] Compared with the prior art, the present invention has the following beneficial effects: the present invention is designed with a dynamic monitoring function, which can comprehensively monitor the vibration frequency, surface anomalies, and operating parameters of automobile engines, and issue real-time alarms for abnormal situations, which can promptly detect potential risks and hidden dangers in the engine, providing strong support for fault prevention and treatment. In addition, a fast cyclic modulation spectrum peak map algorithm is designed through a multi-level decomposition method for detecting and extracting periodic pulse components in engine sound signals in complex acoustic environments. The fast cyclic modulation spectrum peak map algorithm can effectively suppress interference and accurately extract periodic pulse signals through multi-level decomposition and CMSP calculation. The spectral kurtosis method is sensitive to changes in signal-to-noise ratio, while the fast cyclic modulation spectrum peak map algorithm can still maintain a high detection accuracy under low signal-to-noise ratio conditions, thereby improving the accuracy and robustness of acoustic signal fault diagnosis. In addition, the wavelet transform algorithm is used to perform multi-scale decomposition on the collected data, extract fault features, and construct a multi-layer perceptron neural network. This solves the problem that traditional fault diagnosis methods rely on manual experience and have low diagnostic accuracy, improves the automation and intelligence level of fault diagnosis, and through learning and optimization of the neural network algorithm, can accurately identify and predict potential engine faults, providing strong support for engine maintenance and health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a system structure diagram of the present invention;

[0024] Figure 2 This is a module architecture diagram of the dynamic monitoring and recording module of the present invention;

[0025] Figure 3 A diagram showing the module architecture of the user interface and interaction module of the present invention;

[0026] Figure 4 This is a module architecture diagram of the acoustic signal extraction algorithm module of the present invention;

[0027] Figure 5 A module architecture diagram of a neural network module of the present invention;

[0028] Figure 6 is a system flow chart of the present invention;

[0029] Figure 7 Flow chart of the method of the present invention.

[0030] In the figure: 1. Data acquisition module; 11. Vibration frequency acquisition module; 12. Surface anomaly monitoring module; 13. Working condition parameter acquisition module; 14. Acoustic signal acquisition module; 2. Data preprocessing module; 21. Signal filtering module; 22. Data calibration module; 23. Data normalization module; 3. Dynamic monitoring and recording module; 31. Real-time data monitoring module; 32. Data recording and storage module; 33. Abnormal alarm module; 4. User interface and interaction module; 41. Data visualization module; 42. Log recording module; 43. Report generation module; 44. System parameter setting module; 5. Acoustic signal extraction algorithm module; 51. Multiple Level decomposition module; 52. Cyclic modulation spectrum peak calculation module; 53. Fast search strategy module; 54. Periodic pulse signal extraction module; 6. Wavelet transform algorithm module; 61. Multi-scale decomposition module; 62. Fault feature extraction module; 7. Neural network module; 71. Neural network construction module; 72. Input layer node configuration module; 73. Output layer node configuration module; 74. Data set module; 75. Back propagation algorithm training module; 76. Network weight adjustment module; 8. Fault diagnosis and prediction module; 81. Fault identification module; 82. Fault type judgment module; 83. Fault assessment module; 84. Fault prediction module. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0032] Please see the attached Figure 1 -Attached Figure 6The present invention provides an embodiment: a dynamic diagnosis system for automobile maintenance engine faults, comprising a data acquisition module 1, the data acquisition module 1 is connected to a data preprocessing module 2, the data preprocessing module 2 is connected to a dynamic monitoring and recording module 3 and an acoustic signal extraction algorithm module 5, the acoustic signal extraction algorithm module 5 is connected to a wavelet transform algorithm module 6, the wavelet transform algorithm module 6 is connected to a neural network module 7, the neural network module 7 is connected to a fault diagnosis and prediction module 8; the data acquisition module 1 includes a vibration frequency acquisition module 11, a surface anomaly monitoring module 12, an operating condition parameter acquisition module The data preprocessing module 2 includes a signal filtering module 21, a data calibration module 22 and a data normalization module 23. The vibration frequency acquisition module 11 uses a vibration sensor to accurately collect the vibration frequency of the engine under different working conditions, providing basic data for analyzing the engine operating status. The surface abnormality monitoring module 12 uses visual detection and infrared thermal imaging technology to monitor whether there are abnormal wear, cracks or temperature abnormalities on the engine surface. The working condition parameter acquisition module 13 collects the engine's speed, oil pressure, water temperature, oil temperature and other key working condition parameters to fully understand the engine's working status. The signal acquisition module 14 uses an acoustic sensor to collect the sound signal when the engine is running, and provides raw data for subsequent sound signal analysis. The collected raw data is processed by the data preprocessing module 2, specifically the signal filtering module 21 processes the collected raw data, removes high-frequency noise and low-frequency interference, and retains the effective signal. The data calibration module 22 calibrates the collected data to eliminate data deviations caused by factors such as sensor errors and environmental interference. The data normalization module 23 normalizes data of different dimensions to make the data comparable, which is convenient for subsequent data analysis and processing; Dynamic monitoring and recording Module 3 includes a real-time data monitoring module 31, a data recording and storage module 32, and an abnormal alarm module 33. The acoustic signal extraction algorithm module 5 includes a multi-level decomposition module 51, a cyclic modulation spectrum peak calculation module 52, a fast search strategy module 53, and a periodic pulse signal extraction module 54. The real-time data monitoring module 31 is used to continuously track the changes in various parameters of the engine to ensure the real-time and accuracy of the data. The data recording and storage module 32 is used to store the real-time monitored data in a database to facilitate subsequent data analysis and backtracking. The abnormal alarm module 33 is used to set the alarm threshold and issue an alarm in time when abnormal data is detected;The wavelet transform algorithm module 6 includes a multi-scale decomposition module 61 and a fault feature extraction module 62. The neural network module 7 includes a neural network construction module 71, an input layer node configuration module 72, an output layer node configuration module 73, a data set module 74, a back propagation algorithm training module 75 and a network weight adjustment module 76. The multi-scale decomposition module 61 uses the wavelet transform algorithm to perform multi-scale decomposition on the pre-processed data to obtain signals of different frequency bands. The fault feature extraction module 62 identifies feature information related to the fault from these signals, such as frequency changes, amplitude anomalies, etc., and integrates the identified fault features into feature vectors to provide effective data for the input of the neural network. According to support, the neural network construction module 71 in the neural network module 7 constructs a multi-layer perceptron neural network according to the number of extracted fault features, including an input layer, a hidden layer and an output layer, and configures the neural network nodes through the input layer node configuration module 72 and the output layer node configuration module 73. The number of input layer nodes is determined according to the number of extracted fault features, and the output layer nodes correspond to different fault types. The data set module 74 collects a large amount of known fault data as training samples to ensure the diversity and representativeness of the data. The back propagation algorithm training module 75 uses the back propagation algorithm to train the network and adjust the network weight so that the network can accurately identify the fault type. The network weight adjustment module 76 According to the training results and network performance, the network weights are fine-tuned to improve the diagnostic accuracy and robustness of the network; the fault diagnosis and prediction module 8 includes a fault identification module 81, a fault type judgment module 82, a fault assessment module 83 and a fault prediction module 84. The fault identification module 81 uses the trained neural network module 7 to identify the fault of the real-time monitored data and determine whether there is a fault. The fault type judgment module 82 is used to further determine the specific type of fault when a fault is identified and provide clear fault information. The fault assessment module 83 is used to assess the severity of the fault, including the scope of the fault. The fault prediction module 84 is used to determine the fault type based on the current fault. The dynamic monitoring and recording module 3 is connected to the user interface and interaction module 4, which includes a data visualization module 41, a log recording module 42, a report generation module 43, and a system parameter setting module 44. The user interface and interaction module 4 is also connected to the fault diagnosis and prediction module 8. The data visualization module 41 is used to display monitoring data in graphical form, the log recording module 42 is used to record system operation logs to facilitate problem tracking, the report generation module 43 is used to generate diagnostic reports based on monitoring data, and the system parameter setting module 44 is used to set various system parameters.

[0033] Please see the attached Figure 7The present invention provides an embodiment of a method for dynamic diagnosis of automobile engine failure, comprising the steps of: first, data acquisition; second, data preprocessing; third, real-time monitoring and recording; fourth, acoustic signal extraction and analysis; fifth, fault feature extraction; sixth, neural network construction and training; and seventh, fault diagnosis and prediction.

[0034] In the above step 1, the vibration frequency, surface anomaly, operating parameters and acoustic signals of the engine are collected by the data acquisition module 1. The vibration frequency acquisition module 11 uses a vibration sensor to accurately collect the vibration frequency of the engine under different operating conditions, providing basic data for analyzing the engine operating status. The surface anomaly monitoring module 12 uses visual detection and infrared thermal imaging technology to monitor whether there is abnormal wear, cracks or temperature abnormalities on the engine surface. The operating parameter acquisition module 13 collects the engine's speed, oil pressure, water temperature, oil temperature and other key operating parameters to fully understand the engine's working status. The acoustic signal acquisition module 14 uses an acoustic sensor to collect the sound signal when the engine is running, providing raw data for subsequent acoustic signal analysis.

[0035] In the above step 2, the collected raw data is processed by the data preprocessing module 2;

[0036] In the above step 3, the dynamic monitoring and recording module 3 is used to monitor and record the pre-processed data in real time, continuously track the changes in various parameters of the engine, set alarm thresholds, and issue an alarm in time when abnormal data is detected;

[0037] In the above-mentioned step 4, the multi-level decomposition module 51 in the acoustic signal extraction algorithm module 5 decomposes the acoustic signal into signals of different frequency bands through a multi-level decomposition method, and then calculates the cyclic modulation spectrum peak CMSP in each frequency band through the cyclic modulation spectrum peak calculation module 52, and locates the center frequency of the periodic pulse signal by finding the maximum CMSP. In the process, a fast search strategy is adopted by the fast search strategy module 53 to reduce the computational complexity, improve the algorithm efficiency, and ensure real-time performance. Finally, the periodic pulse signal extraction module 54 accurately extracts the engine's periodic pulse signal from a complex acoustic environment based on the located center frequency, providing key information for fault diagnosis.

[0038] In the above step 5, the wavelet transform algorithm module 6 is used to perform multi-scale decomposition on the pre-processed data to extract the fault characteristics;

[0039] In the above step six, the neural network construction module 71 in the neural network module 7 constructs a multilayer perceptron neural network according to the number of extracted fault features, including an input layer, a hidden layer and an output layer. The neural network nodes are configured through the input layer node configuration module 72 and the output layer node configuration module 73. The number of input layer nodes is determined according to the number of extracted fault features, and the output layer nodes correspond to different fault types. The data set module 74 collects a large amount of known fault data as training samples to ensure the diversity and representativeness of the data. The back propagation algorithm training module 75 uses the back propagation algorithm to train the network and adjust the network weights so that the network can accurately identify the fault type. The network weight adjustment module 76 fine-tunes the network weights according to the training results and network performance to improve the diagnostic accuracy and robustness of the network.

[0040] In the above step seven, the trained neural network is used to perform fault diagnosis and prediction on the real-time monitored data.

[0041] Based on the above, the advantages of the present invention are: when the present invention is used to perform dynamic diagnosis of automobile maintenance engine faults, the engine's vibration frequency, surface anomalies, operating parameters and acoustic signals are first collected through the data acquisition module 1. The vibration frequency acquisition module 11 uses a vibration sensor to accurately collect the engine's vibration frequency under different operating conditions, providing basic data for analyzing the engine's operating status. The surface anomaly monitoring module 12 uses visual detection and infrared thermal imaging technology to monitor whether the engine surface has abnormal wear, cracks or temperature anomalies. The operating parameter acquisition module 13 collects the engine's speed, oil pressure, water temperature, oil temperature and other key operating parameters to fully understand the engine's working status. The acoustic signal acquisition module 14 uses an acoustic sensor , collect the sound signal when the engine is running, provide raw data for subsequent sound signal analysis, and then process the collected raw data through the data preprocessing module 2, specifically the signal filtering module 21 processes the collected raw data, removes high-frequency noise and low-frequency interference, and retains the effective signal. The data calibration module 22 calibrates the collected data to eliminate data deviations caused by factors such as sensor errors and environmental interference. The data normalization module 23 normalizes data of different dimensions to make the data comparable, which is convenient for subsequent data analysis and processing. Then, the dynamic monitoring and recording module 3 is used to monitor and record the preprocessed data in real time, continuously track the changes in various parameters of the engine, and set alarm thresholds. When abnormal data is detected, an alarm is issued in time. The real-time data monitoring module 31 is used to continuously track the changes in various parameters of the engine to ensure the real-time and accuracy of the data. The data recording and storage module 32 is used to store the real-time monitored data in the database to facilitate subsequent data analysis and backtracking. The abnormal alarm module 33 is used to set the alarm threshold. When abnormal data is detected, an alarm is issued in time. Then, the multi-level decomposition module 51 in the sound signal extraction algorithm module 5 decomposes the sound signal into signals of different frequency bands through a multi-level decomposition method, and then calculates the cyclic modulation spectrum peak CMSP in each frequency band through the cyclic modulation spectrum peak calculation module 52. The center frequency of the periodic pulse signal is located by finding the maximum CMSP. In the process The fast search strategy module 53 adopts a fast search strategy to reduce computational complexity, improve algorithm efficiency, and ensure real-time performance. Finally, the periodic pulse signal extraction module 54 accurately extracts the engine's periodic pulse signal from a complex acoustic environment based on the central frequency of positioning, providing key information for fault diagnosis. The wavelet transform algorithm module 6 performs multi-scale decomposition on the pre-processed data to extract fault features. Specifically, the multi-scale decomposition module 61 uses the wavelet transform algorithm to perform multi-scale decomposition on the pre-processed data to obtain signals in different frequency bands. The fault feature extraction module 62 identifies feature information related to the fault from these signals, such as frequency changes, amplitude anomalies, etc., and integrates the identified fault features into feature vectors.To provide effective data support for the input of the neural network, the neural network construction module 71 in the neural network module 7 constructs a multi-layer perceptron neural network according to the number of extracted fault features, including an input layer, a hidden layer and an output layer. The neural network nodes are configured through the input layer node configuration module 72 and the output layer node configuration module 73. The number of input layer nodes is determined according to the number of extracted fault features, and the output layer nodes correspond to different fault types. The data set module 74 collects a large amount of known fault data as training samples to ensure the diversity and representativeness of the data. The back propagation algorithm training module 75 uses the back propagation algorithm to train the network and adjust the network weight so that the network can accurately identify the fault type. The network weight adjustment module 76 fine-tunes the network weight according to the training results and network performance to improve the diagnostic accuracy and robustness of the network. Finally, the trained neural network is used to diagnose the fault. The diagnosis and prediction module 8 performs fault diagnosis and prediction on the real-time monitored data. The fault identification module 81 uses the trained neural network module 7 to identify faults on the real-time monitored data and determine whether there is a fault. The fault type judgment module 82 is used to further determine the specific type of fault when a fault is identified and provide clear fault information. The fault assessment module 83 is used to assess the severity of the fault, including the scope of the fault's impact. The fault prediction module 84 is used to predict the development trend and possible occurrence time of the fault based on current data and historical data. Among them, the data visualization module 41 in the user interface and interaction module 4 is used to display the monitoring data in the form of charts. The log recording module 42 is used to record the system operation log to facilitate problem tracking. The report generation module 43 is used to generate a diagnostic report based on the monitoring data. The system parameter setting module 44 is used to set various system parameters.

[0042] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A dynamic diagnosis system for automobile maintenance engine failure, comprising a data acquisition module (1), characterized in that: The data acquisition module (1) is connected to a data preprocessing module (2), the data preprocessing module (2) is connected to a dynamic monitoring and recording module (3) and an acoustic signal extraction algorithm module (5), the acoustic signal extraction algorithm module (5) is connected to a wavelet transform algorithm module (6), the wavelet transform algorithm module (6) is connected to a neural network module (7), and the neural network module (7) is connected to a fault diagnosis and prediction module (8).

2. The automobile maintenance engine fault dynamic diagnosis system according to claim 1, characterized in that: The data acquisition module (1) includes a vibration frequency acquisition module (11), a surface anomaly monitoring module (12), an operating parameter acquisition module (13), and an acoustic signal acquisition module (14); and the data preprocessing module (2) includes a signal filtering module (21), a data calibration module (22), and a data normalization module (23).

3. The automobile maintenance engine fault dynamic diagnosis system according to claim 1, characterized in that: The dynamic monitoring and recording module (3) includes a real-time data monitoring module (31), a data recording and storage module (32), and an abnormal alarm module (33); the acoustic signal extraction algorithm module (5) includes a multi-level decomposition module (51), a cyclic modulation spectrum peak calculation module (52), a fast search strategy module (53), and a periodic pulse signal extraction module (54).

4. The automobile maintenance engine fault dynamic diagnosis system according to claim 1, characterized in that: The wavelet transform algorithm module (6) includes a multi-scale decomposition module (61) and a fault feature extraction module (62), and the neural network module (7) includes a neural network construction module (71), an input layer node configuration module (72), an output layer node configuration module (73), a data set module (74), a back propagation algorithm training module (75) and a network weight adjustment module (76).

5. The automobile maintenance engine fault dynamic diagnosis system according to claim 1, characterized in that: The fault diagnosis and prediction module (8) includes a fault identification module (81), a fault type judgment module (82), a fault assessment module (83) and a fault prediction module (84).

6. The automobile maintenance engine fault dynamic diagnosis system according to claim 3, characterized in that: The dynamic monitoring and recording module (3) is data-connected to a user interface and interaction module (4), the user interface and interaction module (4) comprising a data visualization module (41), a log recording module (42), a report generation module (43) and a system parameter setting module (44), and the user interface and interaction module (4) establishes a data connection with a fault diagnosis and prediction module (8).

7. A method for dynamic diagnosis of automobile engine faults in maintenance, comprising the following steps:

1. data acquisition; 2. data preprocessing; 3. real-time monitoring and recording; 4. acoustic signal extraction and analysis; 5. fault feature extraction; 6. neural network construction and training; and 7. fault diagnosis and prediction. The method is characterized by: In the above step 1, the vibration frequency, surface anomaly, operating parameters and acoustic signals of the engine are collected by the data acquisition module (1); In the above step 2, the collected raw data is processed by the data preprocessing module (2); In the above step 3, the dynamic monitoring and recording module (3) is used to monitor and record the pre-processed data in real time, continuously track the changes in various parameters of the engine, set alarm thresholds, and issue an alarm in time when abnormal data is detected; In the above step 4, the acoustic signal extraction algorithm module (5) accurately extracts the engine's periodic pulse signal from the complex acoustic environment, providing key information for fault diagnosis; In the above step 5, the wavelet transform algorithm module (6) is used to perform multi-scale decomposition on the pre-processed data to extract the fault characteristics; In the above step 6, a multilayer perceptron neural network is constructed according to the number of extracted fault features; In the above step seven, the trained neural network is used to perform fault diagnosis and prediction on the real-time monitored data.

8. The method for dynamic diagnosis of automobile engine failure according to claim 7, characterized in that: In the step 1, the vibration frequency, surface anomaly, operating parameters and acoustic signals of the engine are collected by the data acquisition module (1). The vibration frequency acquisition module (11) uses a vibration sensor to accurately collect the vibration frequency of the engine under different operating conditions, providing basic data for analyzing the engine operating status. The surface anomaly monitoring module (12) uses visual detection and infrared thermal imaging technology to monitor whether the engine surface has abnormal wear, cracks or temperature anomalies. The operating parameter acquisition module (13) collects the engine's speed, oil pressure, water temperature, oil temperature and other key operating parameters to fully understand the engine's operating status. The acoustic signal acquisition module (14) uses an acoustic sensor to collect the acoustic signals of the engine when it is running, providing raw data for subsequent acoustic signal analysis.

9. The method for dynamic diagnosis of automobile engine failure according to claim 7, characterized in that: In the step 4, the multi-level decomposition module (51) in the acoustic signal extraction algorithm module (5) decomposes the acoustic signal into signals of different frequency bands through a multi-level decomposition method, and then calculates the cyclic modulation spectrum peak (CMSP) in each frequency band through the cyclic modulation spectrum peak calculation module (52), and locates the center frequency of the periodic pulse signal by finding the maximum CMSP. In the process, a fast search strategy is adopted by the fast search strategy module (53) to reduce the computational complexity, improve the algorithm efficiency, and ensure real-time performance. Finally, the periodic pulse signal of the engine is accurately extracted from the complex acoustic environment according to the located center frequency through the periodic pulse signal extraction module (54), providing key information for fault diagnosis.

10. The method for dynamic diagnosis of automobile engine failure according to claim 7, characterized in that: In step six, the neural network construction module (71) in the neural network module (7) constructs a multilayer perceptron neural network according to the number of extracted fault features, including an input layer, a hidden layer and an output layer. The neural network nodes are configured through the input layer node configuration module (72) and the output layer node configuration module (73). The number of input layer nodes is determined according to the number of extracted fault features. The output layer nodes correspond to different fault types. The data set module (74) collects a large amount of known fault data as training samples to ensure the diversity and representativeness of the data. The back propagation algorithm training module (75) uses the back propagation algorithm to train the network and adjust the network weight so that the network can accurately identify the fault type. The network weight adjustment module (76) fine-tunes the network weight according to the training results and network performance to improve the diagnostic accuracy and robustness of the network.

Images