Search Results

Radial piston units find several applications in fluid power, offering benefits of low noise and high power density. The capability to generate high pressures makes radial piston pumps suitable for clamping function in machine tools and also to operate presses for sheet metal forming. This study is aimed at developing a comprehensive multidomain simulation tool to model the operation of a rotating cam type radial piston pump, with particular reference to the lubricating gap flow between the pistons and the cylinder block. The model consists of a first module which simulates the main flow through the unit according to a lumped parameter approach. This module evaluates the features of the displacing action accounting for the detailed evaluation of the machine kinematics and for the mechanical dynamics of the check valves used to control the timing for the connection of each piston chamber with the inlet and outlet port.

Hydraulic retarder, as an auxiliary braking device, is widely used in commercial vehicles. Nowadays, the hydraulic retarder’s internal oil is mainly cooled by the coolant circuit directly. It not only aggravates the load of engine cooling system, but also makes the abundant heat energy not be recycled properly. In this study, an independent energy supply device with organic Rankine cycles is applied to solve the problems above. In the structure of this energy supply device, the evaporator’s inlet and outlet is connected in parallel with the oil outlet and inlet of the retarder respectively. A part of oil enters the evaporator to transfer heat with the organic fluid, and the rest of oil enters the oil-water heat exchanger to be cooled by the coolant circuit. According to the different braking conditions of the retarder, the oil temperature in the inlet of the hydraulic retarder can be kept within the proper range through adjusting the oil flow rate into the evaporator properly.

In this paper the nonlinear system identification theory and method using neural networks are presented, the multilayer feedforward networks employed, the backpropagation learning algorithm proposed. The inputs of the networks are consisted of angular velocity and throttle angle, and outputs torque of the engine, finally the comparision of simulation result with that of experiment and other results that embody the effect of system identification are given. Relative studies revealed that the nonlinear system identification for the engine of automated automobiles using neural networks can be effective.