Optimization of Rail Vehicle Operating Speed With Practical Constraints

Abstract

A constrained optimization algorithm to maximize the operating speed of a fifteen degree-of-freedom lateral dynamic model for a passenger railcar subject to random alignment irregularities is presented in this paper. The constraints placed on the optimization problem limit the passenger discomfort, primary and secondary suspension clearance, the wheel slippage, and secondary suspension stroke to practical values while traversing a curve. The optimization results demonstrate that the primary suspension system and the wheel conicity have the most profound influence on maximizing the critical speed where “hunting” begins. The maximum critical speed is insensitive to large variations in secondary yaw stiffness. The secondary lateral stiffness has less effect on the maximum critical speed than the primary lateral stiffness. Thus, the secondary stiffness can be chosen primarily to satisfy passenger ride comfort specifications. The maximum critical speed is quite sensitive to whether the wheel is new, slightly worn, or severely worn.

Optimization of the Lateral Stability of Rail Vehicles

“…Once again [4,5,7], a two-loop method is used for the optimization problem (6). The two-loop method, as shown in Figure 2, is implemented using the MechaGen program [7] (i.e., a GA) as the outer loop algorithm, employing the E04UCF routine (i.e., an SQP) from the NAG (Numerical Algorithms Group) library as the interior loop algorithm.…”

“…When traveling faster than a speed known as the 'critical speed', rail vehicles exhibit this unstable behavior. Since the 1970s, engineers have been designing high speed rail vehicles that can operate at 160 to 480½km=h [4]. Development of a rail vehicle suspension system to operate at these speeds must avoid the serious problem of hunting.…”

“…Cooperider, Hedrick and Cox [4,5] have tackled this task by using an unconstrained optimization method called the Hooke-Jeeves algorithm to optimize the critical speed of a 3 degree of freedom (DOF) rail vehicle model. They also used a constrained optimization method called the M.J.…”

“…Essentially, the GA and SQP are the counterparts of the M.J. Box/Hooke-Jeeves algorithm and the Routh-Hurwitz criteria in the optimization methods described in [4,5]. Eleven design variables, including suspension stiffness and damping, geometric parameters, and inertial property parameters, were optimized [7].…”

Application of SQP and dynamic mode tracking to the determination of the critical speed of rail vehicles

“…Development of a rail vehicle that can operate in the 160-480 km/hr speed regime must avoid the serious problem of hunting (Cox et al, 1978). An effective way to do this is to use optimisation methods to determine a set of suspension parameters that maximise the lateral stability (Cox et al, 1978;McPhee, 2002a, 2002b;Cooperrider et al, 1975;Baumal and McPhee, 2002).…”