Pavement layers are constructed using a combination of materials, of which rock aggregates constitute a larger proportion. Current understanding is that the performance of pavements is dependent on the aggregate shape properties which include form, angularity and surface texture. However, direct and accurate measurements of aggregate shape properties remain a challenge. The current standard test methods used to evaluate aggregate shape properties cannot measure these properties accurately. Among the reasons contributing to the difficulties in the determination of aggregate shape properties is irregular shapes of aggregate particles. Therefore, current research efforts focus on developing accurate, reliable and innovative techniques for evaluation of aggregate shape properties. The work presented in this dissertation contributes to the current innovative research at the Council for Scientific and Industrial Research (CSIR) in South Africa, to automate the measurement of aggregate shape properties. The CSIR’s present research is aimed at improving pavement performance through better materials characterisation, using laser scanning and advanced modelling techniques. The objective of this study was to investigate improved techniques for the determination of aggregate shape properties using analytical and laser scanning techniques. A three-dimensional (3-D) laser scanning device was used for scanning six types of aggregate samples commonly used for construction of pavements in South Africa. The laser scan data were processed to reconstruct 3-D models of the aggregate particles. The models were further analysed to determine the shape properties of the aggregates. Two analysis approaches were used in this study. The first approach used the aggregate’s physical properties (surface area, volume and orthogonal dimensions) measured by using laser scanning technique to compute three different indices to describe the form of aggregates. The computed indices were the sphericity computed by using surface area and volume of an aggregate particle, the sphericity computed by using orthogonal dimensions of an aggregate particle, and the flat and elongated ratio computed by using longest and smallest dimensions of an aggregate particle. The second approach employed a spherical harmonic analysis technique to analyse the aggregate laser scan data to determine aggregate form, angularity and surface texture indices. A MATLABTM code was developed for analysis of laser scan data, using the spherical harmonic analysis technique. The analyses contained in this dissertation indicate that the laser-based aggregate shape indices were able to describe the shape properties of the aggregates studied. Furthermore, good correlations were observed between the spherical harmonic form indices and the form indices determined by using the aggregate’s physical properties. This shows that aggregate laser scanning is a versatile technique for the determination of various indices to describe aggregate shape properties. Further validation of the laser-based technique was achieved by correlating the laser-based aggregate form indices with the results from two current standard tests; the flakiness index and the flat and elongated particles ratio tests. The laser-based form indices correlated linearly with both, the flakiness index and the flat and elongated particles ratio test results. The observed correlations provide an indication of the validity of laser-based aggregate shape indices. It is concluded that the laser based scanning technique could be employed for direct and accurate determination of aggregate shape properties. / Dissertation (MEng)--University of Pretoria, 2013. / gm2013 / Civil Engineering / Unrestricted

One of the major challenges in earth sciences and mineral exploration has been to determine with high accuracy and at a fast rate the elemental composition as well as the general chemistry of a rock sample. Many analytical techniques e.g., scanning electron microscopy (SEM) have been employed in the past with a certain degree of success, but their analyses usually require a lengthy sample preparation and time-consuming measurements which produce results at a much slower rate than techniques whichrequire less or do not require any sample preparation at all. SEM images the surface of a sample by scanning it with a high-energy beam of electrons in a raster scan pattern, where the primary electron beam produced under very low air pressure vacuum scans across the sample by striking it, and a variation of signals produce an image of the surface, or its elemental composition together with energy dispersive X-rays. Alternatively, laser induced breakdown spectrometry (LIBS) and multi energy X-ray transmission (ME-XRT) are non-contact measurement scanning techniques, capable of producing faster results than SEM-EDS which makes them suitable for real time measurements and analyses as they do not slow down the pace of a project being carried out. LIBS is a spectroscopic technique used to characterize and detect materials where a highly energetic laser pulse is focused onto the surfaces of solids, liquids or gases resulting in atomic and molecular species to emit light at specific wavelengths which is collected with a spectrometer and analysed using a computer. Comparably, ME-XRT is a sensor-based sorting technique involving the planar projection of X-ray attenuation of a particle stream, distributed on a fast conveyor belt, where they are scanned and evaluated while passing and an image is recorded by a line scan detector.      Eleven rock samples were analysed in this study. They include four rock type samples: granite, basalt, sandstone, and gneiss, all obtained from Luleå University of Technology (LTU) sample storage and seven ore type samples which include a porphyry Cu sulphide ore, a porphyry Cu oxide ore, a porphyry Cu-Au-Ag ore, an apatite iron ore (AIO), an iron-oxide copper gold ore (IOCG), an orogenic gold ore and a volcanogenic massive sulphide ore (VMS).       The SEM results give a semi-quantitative elemental composition of the rocks, which may be usedto discriminate mineralisation. Energy dispersive X-ray spectroscopy (EDS) maps may be used to identifygeological features and secondary electron (SE) images may be used to understand the topography of the rock samples. The SEM has a low penetration depth rate but produces moderate to high accuracy resultsdepending on the settings and calibrations. It requires a lengthy sample preparation, and its analytical time is often too long for routine industrial application. LIBS results also provide rock elemental compositions similar to the SEM, which may be quantitative if the same spectrometer is used for all elements and calibrated against a standard. It also produces element maps similar to the SEM-EDS maps. LIBS analyses yield high accuracy results but at a low penetration depth. There are no standard calibrations for the LIBS measurements, which limits quantification. LIBS measurements do not require any form of sample preparation. ME-XRT analyses result in rock chemical data portraying a light material fraction (aluminium-like) and a heavy material fraction (iron-like) which may be used to distinguish different rock samples based on the closeness of their effective atomic number Zeff to that of aluminium and iron respectively. It’s analysis also produces low-resolution images of the analysed rock samples. The image resolution is too low to allow interpretation of the data in the context of the structures and textures in the rock samples. It has a higher penetration depth than LIBS and SEM-EDS producing more volumetric data but with a lower accuracy in terms of the amount of information obtained. Only two elements are used for ME-XRT calibration measurements, if many elements of varying atomic numbers could be used, it would have the ability to provide a more reliable data. Samples must have a maximum and minimum thickness; thus, sample preparation is required to regulate the rock thickness.      SEM and LIBS provide element compositions of minerals and element distribution maps required by geologist in their daily activities during exploration and mining. This information can be considered the most useful obtained from all three techniques. However, LIBS analyses are faster, and its maps are of higher quality even at the same resolution as the SEM-EDS. This makes the LIBS preferable for real time measurements and analyses. Geological activities like drill core logging, mine mapping and sampling for grade control all require fast results for project continuity and LIBS is suitable for this purpose as it can keep up with the pace of these activities. SEM analytical technique provides semi-quantitative data which is more accurate than the LIBS data and thus, preferable for usage in research institutions and universities.ME-XRT can reveal information on the internal structures or different rock sample compositions. This makes it a suitable technique in distinguishing ore from waste material especially in iron ore mining and processing where the iron needs to be separated from the siliceous waste and sorting is also required prior to beneficiation to avoid equipment destruction by abrasive quartz. LIBS and ME-XRT analytical techniques complement each other in terms of analytical capabilities as LIBS has a low penetration depthrate but high accuracy results while the ME-XRT has a high penetration depth rate but low accuracy results. They are both fast scanning techniques that can be used for real time measurements and analyses and if their analytical prowess can be improved, the combination of these two fast analytical techniques may enable us to obtain high quality data and may as well be what is needed by geologists in the future.