Abstract
As a crucial output component, a clutch pressure sensor is of great importance on monitoring and controlling a whole transmission system and a whole vehicle status, both of which play important roles in the safety and reliability of a vehicle. With the help of fault diagnosis, the fault state prediction of a pressure sensor is realized, and this lays the foundation for further fault-tolerant control. In this paper, a fault diagnosis method of Dual Clutch Transmission (DCT) is designed. Firstly, a Variable Force Solenoid (VFS) valve model is established. A feed-forward input system is added to correct the first-order inertial link of the sensor on the second step. Finally, the parameters of the established system model are identified by using the measured data of the actual transmission and the Genetic Algorithm (GA). An identified model is then used for designing a fault observer. The constant output faults of 0, 3, and 5 V, pulse fault, and bias fault that enterprises are concerned with are selected to simulate and verify the fault observer under four different operating conditions. The results show that the designed fault observer has great fault diagnosis performance.
Highlights
The clutch pressure sensor, as an important signal output component of the drivetrain, plays a very important role in monitoring and controlling the entire drivetrain and its vehicle
The results show that the designed fault observer has great fault diagnosis performance
When the sensor had a 0 V constant value fault, the fault observer output a signal with a negative amplitude, and the trend of the output was closed to the control current of the solenoid
Summary
The clutch pressure sensor, as an important signal output component of the drivetrain, plays a very important role in monitoring and controlling the entire drivetrain and its vehicle. By of Fault Detection and Isolation(FDI) control based on statistical process monitoring recognizing the patterns of measurement data, the faultssystems were divided into two lever, severity is aimed at raising the possibility of implement in industrial was proposed in [1]. The most informative variables selected tofrom makenoise the decision, and the analyzing fault in different conditions, a fault diagnosis By method was proposed for robustness of thecurrent fault detection and working identification from noise is enhanced. A domain-shared deep residual network was constructed to extract the migration the monitoring data ofmonitoring different pieces equipment; aequipment; domain adaptation regular fault features from the data of of mechanical different pieces of mechanical a domain term constraint was applied in the training of a deep residual network to form a deep adaptation regular term constraint was appliedprocess in the training process of a deep residual network to migration diagnosis model.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have