The current AUSTROADS (1992) pavement design procedure used in New Zealand uses a basic assumption that suggests that a vertical compressive strain applied to any subgrade will cause the same amount of plastic strain to occur irrespective of the type of subgrade. From the work carried out it seems that this assumption appears to be incorrect for volcanic soils.
As stage 1 of this project a literature review was undertaken which indicated that there are a number of models for relating elastic and plastic strain for different materials. A sensitivity analysis was also conducted on the parts of the elastic model used in design and an experimental design developed for testing volcanic soils in stage 2 of this project. In stage 2, five sites were sampled and tested using in-situ CBR, FWD and SASW to determine the in-situ moduli of the pavement. RLT tests were carried out on four of the volcanic soils and one non-volcanic soil obtained from the pavements. Each soil was subjected to one test at each of three nominal axial loads. The test results were fitted to the mathematical models found in the literature.
From the results in stage 2 it was found that the CBR was not a useful indicator of the build up of plastic strain in volcanic soils. The resilient modulus obtained from the RLT tests was found to decrese as the load increased. These moduli were also considerably less than the SASW and CBR predicted moduli but similar to the FWD moduli.
The soil model reviewed in the literature, which produced the 'best-fit' comparision with the RLT data, was found from the model developed by Bayomy and AL-Sanad (1993). Recommendations for further work on determining the relationship between the accumulation of permanent strain in the RLT test and rutting conditions, together with the analysis and testing of different volcanic soils to develop specific failure criteria for each soil are suggested.
Keywords: volcanic subgrades, FWD, subgrade strains, deflections, AUSTROADS, RLT testing, resilient modulus