Published: January 2012 | Category: Activity management , Research programme , Research & reports | Audience: General
Statistical modelling was undertaken to develop a means for reliably predicting the expected in-service skid resistance of any rural section of the New Zealand state highway network surfaced with chip seal.
The measure of slow-speed skid resistance used was the sideways-force coefficient routine inspection machine (SCRIM) coefficients averaged over a 10m length. The statistical modelling was based on data from the 2006–07 high-speed condition survey of the entire New Zealand state highway network, which amounts to a sealed length of 23,113 lane kms. The resulting database contains a total of 976,338 observations allowing identification of statistically significant relationships between the dependent variable (the measured in-service skid resistance) and the independent variables (road geometry, traffic characteristics and aggregate characteristics). One aggregate-related variable investigated was a categorical parameter, representing the name of the quarry from which the aggregate was sourced. This parameter inherently encompasses not only polished stone value but all other important influencing factors such as chip shape, chip hardness, mineralogical properties and crusher type.
The major finding was that the categorical variable 'aggregate source' is a better predictor of in-service skid resistance performance than the numeric variable 'polished stone value'.
Keywords: in-service skid resistance, macrotexture, polished stone value (PSV), predictive modelling, road aggregates, road geometry
