PhD Defense by Ogulcan Canbek

Primary tabs

School of Civil and Environmental Engineering   Ph.D. Thesis Defense Announcement ELUCIDATING THE FRESH AND HARDENED PROPERTIES OF LIMESTONE CALCINED CLAY CEMENTS THROUGH DATA ANALYTICS By Ogulcan Canbek Advisor: Dr. Kimberly Kurtis (CEE) Committee Members: Dr. Newell R. Washburn (Carnegie Mellon), Dr. Yajun Mei( ISYE), Dr. Iris Tien (CEE),  Dr. Susan E. Burns (CEE) Date & Time:  Wednesday June 29, 2022, at 10:00 am Location: https://gatech.zoom.us/j/91309315593?pwd=RFplOTlTUFpFd2p Limestone calcined clay cements (LC3) are a broad class of blended mineral compositions that are alternatives to conventional Portland cement (PC) and are one of the most promising technologies to achieve carbon neutrality in the concrete industry. However, a mechanistic understanding of fresh and hardened properties of LC3-based pastes, mortars, and concrete, as well as empirical design approaches are lacking. This dissertation addresses these knowledge gaps by developing composition-property linkages with the purpose of facilitating the transition of LC3 from the laboratory to practice. Specifically, the influence of LC3 composition on early hydration kinetics, rheological properties, compressive strength development and durability assessed by surface resistivity test is investigated. The compositional design space considers variations in water-to-solids ratio, proportions of constituent materials (PC, calcined clay or “metakaolin” (MK), limestone (LS)), added gypsum content, limestone particle size and superplasticizing admixture dosage. The composition-property linkages are established by combining laboratory data with data analytics approaches including Machine learning (ML). A guiding hypothesis is that the sulfate balance (defined in this dissertation as time difference between the maximum of silicate peak and the sulfate depletion point measured during isothermal calorimetry), influences both the fresh (i.e., rheology) and hardened properties (e.g., compressive strength, surface resistivity) of LC3. To examine this, first, a non-parametric kernel regression technique Nadaraya-Watson (NW) estimator is applied to the heat evolution curves obtained from isothermal calorimetry, allowing quantification of the influence of compositional factors on early hydration kinetics (e.g., slope of silicate peak, sulfate depletion point) in a novel way. Thereafter, linkages between composition and sulfate balance are established first and then the hypothesis of the role of sulfate balance in influencing fresh and hardened properties of LC3 is tested in further chapters. Next, to predict the rheological behavior of LC3, domain knowledge is embedded in ML in the form of five physicochemical predictors, all based on composition. The ML modeling approach helps to elucidate the diversity of mechanisms through which the MK component dominates the rheological behavior of LC3, both directly and through its interactions with the other mineral constituents. Analytical measures (e.g., changes in portlandite and bound water contents over time) show how microstructural development translates to compressive strength and surface resistivity development. For instance, LC3 mortar strength over 28 days of hydration can be accurately predicted not only from its portlandite content over time, but also shows strong correlation with concrete surface resistivity development. Finally, a multi-objective optimization tool is developed to simultaneously predict LC3`s global warming potential and compressive strength development, which are two parameters central in the industrywide shift in cement compositions. Overall, this dissertation provides new


  • Workflow Status: Published
  • Created By: Tatianna Richardson
  • Created: 06/13/2022
  • Modified By: Tatianna Richardson
  • Modified: 06/13/2022