Minimizing Disturbance Detection Time in Hydraulic Systems 2020-01-0263

In a hydraulic system, parameter variation, contamination, and/or operating conditions can lead to instabilities in the pressure response. The resultant erratic pressure profile reduces performance and can lead to hardware damage. Specifically, in a transmission control system, the inability to track pressure commands can result in clutch or variator slip which can cause driveline disturbance and/or hardware damage. A variator is highly sensitive to slip and therefore, it is advantageous to identify such pressure events quickly and take remedial actions. The challenge is to detect the condition in the least amount of time while minimizing false alarms. A Neyman-Pearson and an energy detector (based on auto-correlation) are evaluated for the detection of pressure disturbances. The performance of the detectors is measured in terms of speed of detection and robustness to measurement noise. The implications in terms of computations and memory utilization of implementing the detectors in real-time embedded systems are also discussed. Both simulation and hardware examples are presented. The hardware experiment is performed in a hydraulic system with low damping composed of a solenoid and a regulator valve connected to an electro-hydraulic actuator. From the simulation and hardware experimental data, it is concluded that the auto-correlation detector yields acceptable performance while reducing computation time, and memory usage.