Eleven Instrumented Motorcycle Crash Tests and Development of Updated Motorcycle Impact-Speed Equations

Abstract

Eleven instrumented crash tests were performed as part of the 2016 World Reconstruction Exposition (WREX2016), using seven Harley-Davidson motorcycles and three automobiles. For all tests, the automobile was stationary while the motorcycle was delivered into the vehicle, while upright with tires rolling, at varying speeds. Seven tests were performed at speeds between 30 and 46 mph while four low-speed tests were performed to establish the onset of permanent motorcycle deformation. Data from these tests, and other published testing, was analyzed using available models to determine their accuracy when predicting the impact speed of Harley-Davidson motorcycles. The most accurate model was the Modified Eubanks set of equations introduced in 2009, producing errors with an average of 0.4 mph and a standard deviation (SD) of 4.8 mph. An updated set of Eubanks-style equations were developed adding data published since 2009, and advancing from two equations (pillars/axles and doors/fenders) to four equations (axles, pillars/bumpers, doors, and fenders). When applied to the subject tests, the newly developed set of equations produced an average error of 3.5 mph (SD = 4.3 mph). With respect to all available data (N = 99), the equations produced an average error of 0.1 mph and a standard deviation of 5.8 mph. The errors were also analyzed for each of the four equations developed here, and confidence intervals offered. This research, which represents the first detailed analysis of Harley-Davidson motorcycles’ collision response, indicates they behave in a manner similar to previously tested motorcycles. Further, the equations developed and presented here give accident investigators a refined method for estimating the impact speed of an upright motorcycle, Harley-Davidson or otherwise, having struck an automobile with its front tire. Introduction Analyzing vehicle crush to estimate energy dissipation and thereby calculate vehicle speeds dates back to the 1960s. Jiang et al. [1] provided a thorough summary of the history of speed-from-crush analyses. The earliest work, as well as much of what has come since, focused on passenger cars. The earliest and still most-frequently cited motorcycle testing was conducted by Severy [2], who used seven Honda motorcycles, including: one 90 cc-displacement machine tested at 30mph, five 350 cc units at 20, 30, and 40mph, and one 750 cc machine tested at 30mph. The motorcycles were delivered by dolly such that they struck the side of a stationary 1964 Plymouth sedan in a perpendicular orientation. That research resulted in a linear relationship between approach speed and motorcycle wheelbase reduction. The following equation is the least-squares fit line to that data: S L = ́ + 2 35 8 6 . . eq. 1 Where: S = impact speed (mph). L = motorcycle wheelbase reduction (in). This equation’s coefficient of determination (R-squared value) with the seven data points is in excess of 0.97, indicating a near-perfect fit for this limited dataset with the motorcycle perpendicularly striking the door of a stationary automobile. The correlation between wheelbase change and speed is not nearly as strong in later testing using a wider variety of motorcycles, target vehicles, and different crash configurations. Downloaded from SAE International by Louis Peck, Wednesday, March 28, 2018