Research on engine powerlessness fault diagnosis method based on time-series data mining

This study introduces a novel diagnostic approach for identifying engine powerlessness in commercial vehicles, utilizing time-series data mining technology to replace traditional methods that rely on on-site test drives and incur significant consumption of human and material resources. By analyzing signals from commercial vehicle terminals, we have selected key features closely associated with engine powerlessness faults, such as vehicle speed, acceleration, and the rate of change of throttle opening. We propose the acceleration-intent dual-model fault diagnosis technology scheme, which categorizes data into two groups based on the rate of change of throttle opening: those with acceleration intent and those without. The former utilizes a machine learning model to diagnose the vehicle's power performance, while the latter employs a deep-learning framework with a classification algorithm to classify raw data and detect engine powerlessness. Experimental results demonstrate that this method can identify engine powerlessness faults with high accuracy and specificity, indicating significant potential effectiveness. This research lays a foundation for future work on remote online diagnostic methods for engine-powerlessness faults.