EXPERIMENTAL DESIGN ESTABLISHMENT OF THE STATE HIGHWAY CALIBRATION SECTIONS

Get Complete Project Material File(s) Now! »

Repeatability and Reproducibility

Condition survey equipment accuracy has improved remarkably over the years. However, the mere fact that equipment has a capability of measuring accurately does not necessary means that the equipment system will provide repeatable 1 and reproducible2 results. For this reason, most network data contracts will include some specification that requires a validation/calibration process which the contractor must undertake prior to undertaking network surveys (Transit, 2002). It was therefore expected that the data collection for this research would be even higher than the repeatability and reproducibility criteria required for network surveys.

Rutting Measurement Technology

Rutting has a dual importance in most pavement management systems. Firstly, the physical rut depth is an important parameter from a safety perspective. Aquaplaning may become an issue if the combined factors of cross-fall, longitudinal alignment and rut depth cause water ponding. Furthermore, the rut progression is an important performance measure for roads as it is a strong indicator of the structural behaviour of the pavement under traffic loading. Normally, it is accepted that ruts start to develop as a result of densification within the subgrade once the pavement has reached the design traffic loading. It can also indicate some early failures originating from either the densification of the subgrade or even shear failure in the base or subbase layers (Visser, 1999).

Accuracy of Rutting Measurements Achieved

The rut measurement accuracy and repeatability obtained during this research was acceptable for the given objectives and assumptions made during the onset of the research (refer to Section 4.6). It was also observed that the annual rut change was relatively small but correlated with the expected values during the onset of the research. Figure 4.11 illustrates the distribution of rut changes for the first three survey periods of this research. All results are plotted in these figures, and the negative values represent sections which have been maintained, and some which showed a minor rut improvement which was within the measurement tolerance.

Performance Based Specifications

For the data collection of the State Highway LTPP sections, as noted in Chapter 3, Transit decided to appoint a private contractor. Hence, the data collection contract had to comply with Competitive Pricing Procedure (CPP) requirements (Transfund, 1997). The contractual requirements had an additional benefit for this LTPP research, in the sense that precision, repeatability and reproducibility requirements had to be specified in detail. This approach is different to many other situations where researchers are often more focused on prescribing the method of data collection. The additional benefits were realised in that the tender had to specify exactly what the desired outcome had to be, whilst this aspect is not always well thought through in other research projects. Therefore, research projects are also constrained with regards to the availability and/or the affordability of equipment.

Existing Model Format

According to the HDM definition, crack initiation occurs when a surface displays cracks on more than 0.5% of its area (Watanatada, et al. 1987). The cracked area is calculated by multiplying the length of the crack by the width of affected area (for line cracks the effected width is assumed to be 0.5m). Separate crack initiation model forms were developed in HDM-4 for stabilised and granular bases, and for original surfaces and resurfaced surfaces. The majority of New Zealand roads fall in the granular base category, as most New Zealand pavements are only lightly stabilised in contrast to global practice.

Calibration Methodology of the World Bank HDM-4 Models

The main objective of Level 2 calibration is to adjust the calibration coefficients so that the predicted performance compares well with the observed performance. For example, in an area with high rainfall it would be expected that pavements would deteriorate faster than similar pavements in a dry area. In such a scenario, the wet area would have a crack initiation calibration coefficient of less than 1.0, which means that pavements would crack faster than pavements modelled in the original study area.

Table of Contents :

• 1 INTRODUCTION
  • 1.1 The Context of Pavement Deterioration Models
  • 1.2 The Historical Development of Pavement Modelling in New Zealand
  • 1.3 Problem Statement
  • 1.4 Objectives of the Research
  • 1.5 Scope and Structure of the Research Report
• 2 LITERATURE REVIEW
  • 2.1 Literature Review LongTerm Performance Studies
  • 2.2 World Bank HDMIII LTPP Studies in Kenya and Brazil
  • 2.3 North Americas – SHRP Study
  • 2.4 Australia – Development of New Pavement Models
  • 2.5 Australia – Calibration of HDM4 Pavement Models
  • 2.6 South African (Gautrans) HDMIII and HDM4 Calibration Studies
  • 2.7 Guidance for the Research
• 3 EXPERIMENTAL DESIGN ESTABLISHMENT OF THE STATE HIGHWAY CALIBRATION SECTIONS
  • 3.1 Introduction
  • 3.2 Climatic Stratification
  • 3.3 Traffic/Loading
  • 3.4 Pavement Strength/Pavement Types
  • 3.5 Condition / Age
  • 3.6 Experimental Design for this research
  • 3.7 Site Identification and Selection Criteria
  • 3.8 Statistical Summary of LTPP Sections Established
• 4 LTPP DATA COLLECTION
  • 4.1 Introduction
  • 4.2 Theoretical Definitions and Considerations Related to the Data Collection
  • 4.3 Roughness Measurements
  • 4.4 Rutting Measurements
  • 4.5 Visual Surveys
  • 4.6 Survey Specifications
  • 4.7 Discussion on Appropriateness of Data Collection Regime
• 5 PREDICTING CRACK INITIATION
  • 5.1 Introduction
  • 5.2 Calibration of the HDM4 Model
  • 5.3 Adjustment of HDM4 Default Model Coefficients
  • 5.4 Development of an Alternative Crack Initiation Model
  • 5.5 Discussion
  • 5.6 Crack Initiation Summary
• 6 PREDICTING RUT PROGRESSION
  • 6.1 Introduction
  • 6.2 Analysis Objectives and Data Use
  • 6.3 HDM Rut Models
  • 6.4 Predicting Initial Densification
  • 6.5 Rut Progression
  • 6.6 Accelerated Rutting
  • 6.7 Rut Progression Summary
• 7 THIS RESEARCH IN CONTEXT
  • 7.1 Purpose of this Chapter
  • 7.2 LTPP Experimental Design
  • 7.3 Data Collection
  • 7.4 New Pavement Prediction Models
  • 7.5 Past and Future Use of the LTPP Data from a National Perspective
• 8 CONCLUSIONS AND RECOMMENDATIONS
  • 8.1 Conclusions
  • 8.2 Recommendations – Models Developed
  • 8.3 Further Work
  • 8.4 Lessons Learnt from this Research
• 9 REFERENCE LIST

GET THE COMPLETE PROJECT
The Development of Pavement Deterioration Models on the State Highway Network of New Zealand