Concrete admixtures are constantly used in construction projects, and these admixtures can be used to increase/decrease the setting time, improve workability, enhance frost and sulfate resistance, and to help control strength development. The quality of admixtures play a crucial role in altering fresh concrete properties. To streamline such quality control process for regular inspections, two key procedures are necessary: (1) collection and maintenance of the baseline approved admixtures quality, and (2) usage of sound analysis tools and procedures for quality verification of newly supplied materials. However, relying on manual comparison by the naked eye makes the identification process very tedious and inaccurate due to similar patterns of admixture infrared spectrophotometry (IR) signatures. Therefore, an efficient and standardized system is necessary for an accurate quality control process. In this study, advanced quality control models are investigated and proposed by utilizing the pattern-recognition methodology using artificial neural networks (ANNs) by utilizing data-driven and self-adaptive functions and other advanced machine-learning techniques for the pattern classification function. The feasibility of the proposed models were evaluated for the automatic quality control process. To identify a mixture's chemical and physical properties, energy absorption is measured at each wave number through the amount of transmitted infrared light. Developed pattern-recognition ANNs and other machine-learning models have shown their efficiency in learning and identifying the admixtures' IR signature spectra. Hence, the proposed advanced quality control models can be a very useful tool to determine the admixtures' quality accurately and quickly and eventually to guarantee their intended performance in altering the concrete's properties.

Advanced Quality Control Models for Concrete Admixtures

Gregori A.;
2019-01-01

Abstract

Concrete admixtures are constantly used in construction projects, and these admixtures can be used to increase/decrease the setting time, improve workability, enhance frost and sulfate resistance, and to help control strength development. The quality of admixtures play a crucial role in altering fresh concrete properties. To streamline such quality control process for regular inspections, two key procedures are necessary: (1) collection and maintenance of the baseline approved admixtures quality, and (2) usage of sound analysis tools and procedures for quality verification of newly supplied materials. However, relying on manual comparison by the naked eye makes the identification process very tedious and inaccurate due to similar patterns of admixture infrared spectrophotometry (IR) signatures. Therefore, an efficient and standardized system is necessary for an accurate quality control process. In this study, advanced quality control models are investigated and proposed by utilizing the pattern-recognition methodology using artificial neural networks (ANNs) by utilizing data-driven and self-adaptive functions and other advanced machine-learning techniques for the pattern classification function. The feasibility of the proposed models were evaluated for the automatic quality control process. To identify a mixture's chemical and physical properties, energy absorption is measured at each wave number through the amount of transmitted infrared light. Developed pattern-recognition ANNs and other machine-learning models have shown their efficiency in learning and identifying the admixtures' IR signature spectra. Hence, the proposed advanced quality control models can be a very useful tool to determine the admixtures' quality accurately and quickly and eventually to guarantee their intended performance in altering the concrete's properties.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11697/141380
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 3
  • ???jsp.display-item.citation.isi??? 3
social impact