Factors Influencing the InService

1

• Factors Influencing the In-Service
• Skid Resistance Performance of Aggregates
• Peter Cenek1
• Robert Davies2
• Robert Henderson1
• 1Opus International Consultants
• 2Statistics Research Associates

2
Presentation Outline

• Preliminary findings from New Zealand Transport
  Agency
• Research Contract LTR 0021
• Selection of Aggregates for Skid Resistance

3
Hypothesis to be Tested

• The categorical parameter
• Quarry from where the roading aggregate is
  sourced
• is a better predictor of
• in-service skid resistance performance
• than the Polished Stone Value (PSV) of the
  aggregate.

4
Need for Research

• NZTA presently spends 4.5 5 million per annum
  on SCRIM sealing.
• It was expected that 5 years after implementation
  of T/10, SCRIM sealing would reduce to 1 million
  per annum (2003-04 onwards).
• Is the over expenditure due to
• Incorrect recording of resealing activity or
• Over-reliance on aggregate PSV to achieve
  required in-service skid resistance?

5
Basis of Specification for PSV Requirements

• The Szatkowski Hosking equationis used in both
  the UK and NZ as the basis for specifying the
  PSV of aggregates employed in the construction
  of new road surfaces.
• SC 0.024 0.663 ? 10-4CVD 0.01PSV (r20.83)
• where SC SCRIM Coefficient CVD
  Commercial vehicles per lane per day PSV
  Polished Stone Value

6
Subsequent Studies

• UK (Roe Hartshorne, 1998) and NZ (Cenek et al.,
  2004) findings suggest PSV based equations do
  not reflect on-road performance well (r2lt0.1).

Source Roe Hartshorne (1998)
7
Methodology

• Express ESC in terms of the independent variables
  using linear regression.
• Regression procedure required that models the
  random structure one might expect.
• Two levels of randomness supposed
• ESC measurements
• Surface layer information from RAMM.
• Provides more realistic tests of significance and
  confidence intervals than simple regression
  analysis.

8
Road Surface Types Considered

• 1. A single coat seal (shown as a reseal)
• 2. A two-coat seal (shown as a first coat)

Uniformly sized chip
Binder
Old chipseal
Basecourse
Second application of binder, bitumen-coated
larger chips are visible from above
Second (smaller) chip
First (larger) chip
First application of binder
Basecourse
Source Chipsealing in New Zealand (2005)
9
Dataset

• Confined to 2004 SCRIM Survey of the Northland
  (723 km) and Napier (815 km) Management Areas.
• The 10 m sections analysed represented 14 of
  NZs State Highway Network.
• Observations excluded
• Texture depth lt 0.5 mm MPD
• ADT lt 100 vpd
• Absolute Horizontal Curvature lt 10m
• Age lt 2 years

10
Data Sources

• NZTA Road Asset Information System RAMM
• 1. 2004 SCRIM Survey of State Highway Network
• 2. State Highway Traffic Monitoring
• 3. Surfacing Tables

11
SCRIM Data

• Road Geometry
• Horizontal Curvature m 10m intervals
• Gradient 10m intervals
• Cross - fall 10m intervals
• Road Condition
• Lane Roughness IRI m/km 20m intervals
• Rut Depth mm 20m intervals
• Texture Depth mm MPD 10m intervals
• Skid Resistance SCRIM Coeff. 10m intervals

12
SCRIM Operated by WDM (UK) Ltd.
13
Statistically Significant Predictor Variables

• Macrotexture
• Horizontal Curvature
• Longitudinal Gradient
• Skid Resistance Site Category
• Daily Traffic (ADT)
• Seal Age
• Seal Type
• Speed Environment
• Quarry Source of Aggregate

14
Effects of Quantitative Predictor Variables
Macrotexture
15
Effects of Quantitative Predictor Variables
Reciprocal of Horizontal Curvature
16
Effects of Quantitative Predictor Variables
Gradient
17
Effects of Quantitative Predictor Variables
ADT
18
Effects of Quantitative Predictor Variables
Age
19
Effect of PSV

• Analysis of variance where terms are added
  sequentially.
• Aggregate source added after PSV
• Both statistically very significant.
• PSV added after aggregate source
• Only aggregate source statistically significant.
• ð Aggregate source has more predictive power than
  PSV

20
Ranking of Northland and Napier Sources
21
Concluding Remark

• There is a strong case to use statistical
  modelling
• to complement PSV test results
• when ranking suppliers of surfacing aggregates.

22
Thank You for Your Interest