Nonstationary coherence characteristics of dual track road profile data

The left to right wheel-path coherence function was proven to be nonstationary. The variations in the coherence could be described using the distribution of a single arbirary exponent of an exponential curve-fit. The distribution of the expoent was shown to be independent of overall road roughness. [Display omitted] •Analysed left–right wheel-path coherence function of 400 km of asphalted roads.•Assessed all published coherence models to identify the most appropriate.•Developed new approach to apply coherence model to nonstationary roads.•Showed that wheel path coherence function is independent of road roughness.•Established a new method for statistically describing road roughness distributions. The ability to accurately simulate the vibratory motion of transport vehicles is of great importance when designing vehicle components and product containment systems. Direct measurement and analysis of the vibrations is not always practical and laboratory testing using synthesized road elevation data is a common alternative, as is numerical simulation. However, no technique exists to generate realistic nonstationary dual track road elevation data. This research focuses on uncovering statistical distributions that describe the nonstationary relationships between the left and right wheel-paths. Analysis of the short-time (nonstationary) coherence functions and instantaneous International Roughness Index (IRI) of measured road profile data provided distributions which describe variations in left to right wheel-path correlation and roughness variations for both tracks. The resulting distributions can be described with a three-parameter Weibull distribution and can be adopted to generate nonstationary dual wheel-path profile data that can be used to excite numerical vehicle models and physical vehicles via multi-axis simulators.