Maybelle Liang

Rock strength assessment based on regression tree technique

Abstract Uniaxial compressive strength (UCS) is one of the most important parameters for investigation of rock behaviour in civil and mining engineering applications. The direct method to determine UCS is time consuming and expensive in the laboratory. Therefore, indirect estimation of UCS values using other rock index tests is of interest. In this study, extensive laboratory tests including density test, Schmidt hammer test, point load strength test and UCS test were conducted on 106 samples of sandstone which were taken from three sites in Malaysia. Based on the laboratory results, some new equations with acceptable reliability were developed to predict UCS using simple regression analysis. Additionally, results of simple regression analysis show that there is a need to propose UCS predictive models by multiple inputs. Therefore, considering the same laboratory results, multiple regression (MR) and regression tree (RT) models were also performed. To evaluate performance prediction of the developed models, several performance indices, i.e. coefficient of determination (R2), variance account for and root mean squared error were examined. The results indicated that the RT model can predict UCS with higher performance capacity compared to MR technique. R2 values of 0.857 and 0.801 for training and testing datasets, respectively, suggests the superiority of the RT model in predicting UCS, while these values are obtained as 0.754 and 0.770 for MR model, respectively.

Performance evaluation of existing surface excavation assessment methods on weathered sedimentary rock

This paper is intended as a contribution to the discussion of the overview of existing excavatability assessments and their reliability in assessing the state of tropically weathered rocks, specifically in Nusajaya, Johor, Malaysia. For this purpose, case studies were conducted on sedimentary rock masses of the Jurong formation, located in Nusajaya, Johor. The existing available rock mass excavatability assessments were first critically reviewed. Three case studies were then conducted in order to assess the performance of surface excavatability in these sedimentary rock masses by practical excavation tests on site. Comparisons were then made between the assessments of those assessors and that obtained through the practical excavation test. The conclusions made in the paper about the results of the study show many divergent opinions among the assessors. It emphasizes the limitations of existing available assessments in predicting the excavatability of tropically weathered rock masses. Further complication may arise in sedimentary rock masses where different layers of rock are interbedding, producing different weathering profiles. In view of these, a simpler graphical method, which considers only two parameters, namely point load strength and spacing of surface discontinuities, was introduced to overcome the problems, thus might be misjudged during early excavation assessment. This subjectivity becomes more important as the heterogeneity of a rock mass increases with increasing interbedding and inherent anisotropy of the rock material, giving rise to quite different evaluations, even for a rock with the same state of weathering. Most of the existing excavatability assessment methods are found to be less reliable because nature, lithology and weathering state at various rock mass layers is not being taken into account. A more practical and appropriate excavatability assessment method is thus required to economically assess the tropically weathered sedimentary rock during the preliminary stage.

An excavatability classification system for surface excavation in sedimentary rocks

This paper describes the development of an excavatability classification system for surface excavation of weathered sedimentary rock based on practical excavations and statistical analysis. Extensive field and laboratory studies were undertaken on three sedimentary rock sites with a range of weathering and strength conditions in the Iskandar region, South Johore, Malaysia. Rock mass descriptions were made in the field and rock material tests were carried out in the laboratory. Statistical analysis was used to study the correlation between the rock parameters and the productivity of excavation. A classification rating system based on the analysis results is proposed. The proposed classification system incorporates rock parameters such as point load strength (Is50), Schmidt Rebound hammer value (R), uniaxial compressive strength (UCS), density of rock (γ), indirect tensile strength (ITS), second-cycle of slake durability index (Id2), joint number (Jn), joint spacing (Js), joint length (JL), joint direction (Jd), existence of iron pan (IP), moisture content (Mc) and existence of bedding. The proposed model was then verified against further excavations on another sedimentary rock site. It was found that the productivities recorded correlated well with the proposed system when tested on a fourth site.

Artificial neural network models for FRP-repaired concrete subjected to pre-damaged effects

Confining damaged concrete columns using fibre-reinforced concrete (FRP) has proven to be effective in restoring strength and ductility. However, extensive experimental tests are generally required to fully understand the behaviour of such columns. This paper proposes the artificial neural networks (ANNs) models to simulate the FRP-repaired concrete subjected to pre-damaged loading. The models were developed based on two databases which contained the experimental results of 102 and 68 specimens for restored strength and strain, respectively. The proposed models agreed well with testing data with a general correlation factor of more than 97%. Subsequently, simplified equations in designing the restored strength and strain of FRP-repaired columns were proposed based on the trained ANN models. The proposed equations are simple but reasonably accurate and could be used directly in the design of such columns. The accuracy of the proposed equations is due to the incorporation of most affecting factors such as pre-damaged level, concrete compressive strength, confining pressure and ultimate confined concrete strength.