Simple Predictors of the Performance of A-trains 932995

Figures of merit describing the performance qualities of multiple-trailer vehicle combinations (for example, rearward amplification) are usually determined from either full-scale vehicle testing or computer simulation analysis. Either method is expensive and time consuming, and restricted in practice to organizations with specialized equipment and engineering skills. One goal of a recent study, conducted by the University of Michigan Transportation Research Institute and sponsored by the Federal Highway Administration, was to use basic vehicle properties to develop simple formulations for estimating the performance qualities of multiple-trailer vehicle combinations.

Several hundred computer simulation runs were made using UMTRI's Yaw/Roll program. Five common double-trailer vehicle configurations (defined by trailer lengths and axle configurations) were studied. Each of the five vehicles was subject to fifteen parameter variations. These variations involved geometry, tire, or load characteristics. Simulations of four types of maneuvers were conducted. The simulation results were reduced to generate performance figures of merit and linear regression techniques were used to develop formulas for predicting performance as a function of vehicle properties. For several performance measures, formulas yielding surprisingly good correlation between predicted and measured performance were found. As could be expected, tire cornering stiffness, trailer wheelbase, and vehicle qualities related to roll were most important in predicting performance.