Scholarlink

The synthesized aggregate manufacturing involves the geopolymerization technology and the direct pressure application approach to obtain coarse aggregates. This current study combined industrial-grade FAS, GGBS, Fine Aggregate (FAgg), Sodium Hydroxide (NaOH), and Sodium Silicate (Na2SiO3) to create a geopolymer mortar. Initially, geopolymer mortar cubes of 150 mm were created by applying geopolymer technique. Subsequently, these cubes were broken down into coarse stones having angular forms at 7 days. The artificially synthesized aggregates were employed as geopolymer coarse aggregates (GCA) in lieu of conventional coarse aggregates for the production of geopolymer aggregate concrete (GPAC). Since experimentation alone cannot solve the problem of data scarcity, a feed-forward network is used to address this challenge and is modeled as a regression problem. The issue of data insufficiency was overcome by creating 1000 synthetic data points with the Conditional Tabular Generative Adversarial Network (CTGAN) algorithm. The proposed approach achieved a mean squared logarithmic loss of 0.0277 and an R2 score of 0.5491.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction of Compressive Strength in Geopolymer Coarse Aggregates by Synthesizing Data Using CTGAN

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • K. Krishna Bhavani Siram,
  • Dhanya Sathyan,
  • B. Premjith

摘要

The synthesized aggregate manufacturing involves the geopolymerization technology and the direct pressure application approach to obtain coarse aggregates. This current study combined industrial-grade FAS, GGBS, Fine Aggregate (FAgg), Sodium Hydroxide (NaOH), and Sodium Silicate (Na2SiO3) to create a geopolymer mortar. Initially, geopolymer mortar cubes of 150 mm were created by applying geopolymer technique. Subsequently, these cubes were broken down into coarse stones having angular forms at 7 days. The artificially synthesized aggregates were employed as geopolymer coarse aggregates (GCA) in lieu of conventional coarse aggregates for the production of geopolymer aggregate concrete (GPAC). Since experimentation alone cannot solve the problem of data scarcity, a feed-forward network is used to address this challenge and is modeled as a regression problem. The issue of data insufficiency was overcome by creating 1000 synthetic data points with the Conditional Tabular Generative Adversarial Network (CTGAN) algorithm. The proposed approach achieved a mean squared logarithmic loss of 0.0277 and an R2 score of 0.5491.