Automobile fault diagnosis method and system

Through the federal learning framework, the automotive fault diagnosis method combining exhaust gas and environmental parameters is dynamically calibrated λ value and exhaust gas threshold, and the engine combustion state is inverted, solving the problem of insufficient accuracy of traditional diagnostic technology under complex operating conditions, and achieving high-precision fault identification and adaptive diagnosis.

CN120429759AActive Publication Date: 2025-08-05JILIN COMM POLYTECHNIC +2
4 Cites 1 Cited by

Patent Information

Application Number
CN202510920653.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-04
Publication Date
2025-08-05
Estimated Expiration
2045-07-04

AI Technical Summary

Technical Problem

Traditional automotive fault diagnosis technology has insufficient fault feature recognition accuracy under multi-parameter dynamic coupling and environmental interference, and cannot adapt to complex working conditions, resulting in high missed detection rate and outstanding risk of misjudgment.

Method used

The vehicle fault diagnosis method based on the federal learning framework is adopted, combining exhaust gas component data and local environmental parameters, and dynamically calibrates the λ value and exhaust gas component threshold through physical information neural network and CNN-Transformer classifier to generate normalized exhaust gas data after environmental compensation, invert the combustion state parameters in the engine cylinder, and output the fault probability distribution.

Benefits of technology

It significantly improves the accuracy of fault diagnosis, realizes adaptive diagnosis under different environmental conditions, and enhances the coordinated analysis of exhaust gas timing characteristics and long-term dependence relationships in engine operating conditions.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses an automobile fault diagnosis method and system, and relates to the technical field of fault diagnosis, and the method comprises the steps: inputting tail gas component data and local environment parameters into a pre-trained global reference model based on a federated learning framework, dynamically calibrating a lambda value and a tail gas component threshold value, and generating environment-compensated normalized tail gas data; the method comprises the following steps: extracting tail gas time sequence local features through a convolutional layer, modeling a long-period dependency relationship of OBD-II data through a Transform encoder, outputting combustion fault probability distribution and mechanical fault probability distribution, encrypting and uploading the distribution to a federated server, and updating a global reference model. According to the method, the tail gas component data, the environmental parameters and the OBD-II data are combined and embedded into the combustion reaction kinetic equation, so that the combustion state in the engine cylinder can be reflected more accurately, and the accuracy of fault diagnosis is remarkably improved.
Need to check novelty before this filing date? Find Prior Art