Yoshiaki Saito

A method of predicting the time to fill the hydraulic pipeline considering variance

The oil filled in the hydraulic pipe connected to the clutch piston influences the clutch engagement in the transmission of construction machinery. In order to improve shift control, we need to catch the precision time to fill the pipe with oil (hereinafter called filling time). However, oil dynamics have nonlinearity which makes it difficult to predict the filling time with high accuracy. Moreover, the filling time actually has the variance so estimating the filling time including the variance is necessary for considering the control parameter quantitatively. Hence, the purpose of this research is to consider the method of calculating the amount of oil filled in the hydraulic pipe and predicting the filling time including variance. We use neural network model for creating the model estimating the flow rate model and propose the method of expressing the variance of the filling time with particles.