Fault Diagnosis of Automotive Engine Intake System Vacuum Leaks Based on GA and RBF

Fault diagnosis of automotive engines based on fault codes, waveforms, and data streams can significantly improve diagnostic efficiency. However, in practical vehicle use, some specific faults are not well addressed. This paper collects automotive engine exhaust gas data and uses RBF neural networks classification model and the RBF neural network classification model optimized by genetic algorithms to determine the presence of specific faults. The results show that the RBF network achieves a detection accuracy of 93.33% for intake system vacuum leaks, while the RBF neural network with genetic algorithm optimization reaches a detection accuracy of 100% for intake system vacuum leaks.