Strength prediction and mix design procedures for geopolymer and alkali-activated cement mortars comprising a wide range of environmentally responsible binder systems. Issue 3 (3rd July 2019)

• Record Type: Journal Article
• Authors: Oakes, L.; Magee, B.; McIlhagger, A.; McCartney, M.
• Is Part Of: Journal of structural integrity and maintenance. Volume 4:Issue 3(2019)
• Page Start: 135
• Page End: 143
• Publication Date: 2019-07-03
• Subjects: Geopolymer -- alkali activated -- metakaolin -- industrial wastes -- mixture design -- performance prediction; Structural stability -- Periodicals; Safety factor in engineering -- Periodicals; Structural engineering -- Periodicals; 624.171
• DOI: 10.1080/24705314.2019.1622189
• Languages: English
• ISSNs: 2470-5314

Abstract

ABSTRACT: This research presents a mix design methodology for geopolymer (GP) and alkali activated (AA) mortars based on metakaolin and industrial waste products activated using potassium silicate. The aim is to enable a wide range of mix designs to be specified to given compressive strength, consistency and environmental footprints to facilitate their adoption as a Portland cement (PC) alternatives in the construction industry in applications such as fibre-reinforced building cladding systems. The impact of the work is timely as literature quantifying the effects of mix parameters on broad families of GP and AA materials are limited, and no standardized performance-based methodology exists. Initially, effects of binder composition on mechanical and environmental properties are presented for a standard GP mix design using contoured ternary plots for a range of material blends. Next, effects of altering mixture parameters such as liquid/solid, silica/alumina and activator/binder ratios are quantified for three selected binder compositions before a preliminary mix design methodology is presented allowing initial selection of mixture proportions. Finally, correlation analysis is used to identify multiple mix variables strongly correlating with strength, and regression modeling used to present predictive tools with average errors <6%.