Powder packing optimisation for clinker reduction in concrete
Master Thesis
2018
Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Department
License
Series
Abstract
Globally, concrete is the most used construction material. Its embodied energy is relatively low, yet due to the vast quantities that are produced annually, it has substantial greenhouse gas (GHG) emissions associated with it. Of the concrete constituents, the manufacture of clinker - the basis of all conventional cements - contributes the most significant emissions. Therefore, to reduce the emissions associated with concrete manufacture, there has been extensive research into how clinker content can be reduced without compromising desired concrete properties. Existing methods for clinker reduction have, however, only allowed clinker replacement to a limited extent. This research investigated the more efficient use of clinker to minimise clinker content required to achieve desired mechanical and durability properties of concrete. The optimisation of powder (materials < 125 µm) packing, using filler materials with varying fineness, was identified to potentially increase clinker efficiency. The optimisation undertaken was the maximisation of powder packing density but without adversely affecting workability. The investigation entailed the application of analytical particle packing density models as well as experimental investigation. Two particle packing models, the Compaction Interaction Packing Model (CIPM) and the Modified Andreasen and Andersen Curve (MAAC) were applied. Various methods for determining the packing density of powder combinations were investigated which informed the use of the mixing energy test to provide experimental packing density data for the modelling procedures. The CIPM was used to optimise the powder phases of concrete as it incorporated the effect of surface forces on powder packing and the MAAC was used to complete the optimisation of fine and coarse aggregate materials. It was necessary to calibrate the CIPM through the selection of various model constants, based on the minimisation of the average error associated with predicted packing density. Despite the incorporation of surface force effects, the CIPM did not predict the trend in packing density observed for various experimental powder combinations with consistent accuracy. Combinations of cement with limestone of high and low fineness (relative to cement) were most accurately predicted but combinations with limestones of similar fineness to cement were less accurate. It was therefore apparent that the model inadequately accounted for the effects of varying particle size and the corresponding influence of surface forces on these particles. However, for practicality, model constants which minimised overall error were used to determine powder combinations enabling maximum packing density for use in optimised concrete mix design. Concrete mixes were designed in 2 phases. Initially water content was fixed, and limestone content was successively increased to 40 vol. % (Phase 1). Despite the formation of mixtures according to maximum packing density, the results showed that optimisation of packing density with a fixed water content was insufficient to reduce clinker content without adversely affecting compressive strength. However, workability was maintained without excessive superplasticiser (SP) dosage and oxygen permeability, water sorptivity and accelerated drying shrinkage were either improved or not adversely affected. This was attributed to the ability of fine fillers to prevent interconnectivity of the pore structure and the decreased volume of gel hydration products leading to reduced drying shrinkage. Compressive strength was tested for a binary (cement/limestone) and ternary (cement /limestone/fly ash (FA)) binder blend for Phase 2 in conjunction with a substantially reduced water content. Workability was adversely affected and both mixes required high SP doses, however, the FA blend required a relatively lower dose. Compressive strength was again decreased relative to the reference mix but when comparing Phase 1 and 2 mixes with predicted strength for equivalent w/c ratios, compressive strength was relatively unchanged, inferring little benefit of packing optimisation. However, binder efficiency indices (‘bi’) (between 5.3 and 6.9 kg/m3 /MPa) were reduced relative to data from previous investigations with similar strength class (between 10 to 20 kg/m3 /MPa), inferring increased binder performance. Powder packing optimisation thereby has the potential to enable clinker reduction, particularly for lower strength grade concrete, without adversely affecting compressive strength. Furthermore, the relatively unaffected durability indicators portray the beneficial effects of powder packing optimisation on increasing the impenetrability of concrete microstructure and it potential use in applications where durability is of importance. These findings also pointed to further possible reductions in the binder efficiency index below 5 kg/m3 /MPa if water content is further reduced (to maintain low water: cement ratio) and reactive SCMs are incorporated. However, further investigation and understanding of the fundamentals of powder packing is necessary to achieve a fully predictive process of low-clinker concrete mix design that can be universally applicable.
Description
Reference:
Holmes, M. 2018. Powder packing optimisation for clinker reduction in concrete. University of Cape Town.