Optimization design for alkali-activated slag-fly ash geopolymer concrete based on artificial intelligence considering compressive strength, cost, and carbon emission

Abstract

Traditional mix proportion design methods are difficult to consider the multiple objectives, while the design of economical and environmentally friendly alkali-activated slag-fly ash geopolymer concrete on the premise of guaranteeing compressive strength is an essential and meaningful task. The novelty of this paper is that the optimization design model for alkali-activated slag-fly ash geopolymer concrete considering 28 days compressive strength, cost, and carbon emission was developed based on machine learning and Particle Swarm Optimization (PSO) algorithm. The results show that Random Forest (RF), GB, and Back Propagation Neural Network (BPNN) can achieve a good prediction effect for compressive strength with R2 over 0.85 and 0.70 for training and testing set. Slag and sodium hydroxide (NaOH) contents have remarkable effects on the compressive strength of alkali-activated slag-fly ash geopolymer concrete. The addition of slag is beneficial to enhancement of compressive strength, and the relatively optimal coarse and fine aggregate contents are 1200 kg/m3 and 750 kg/m3. It is noted that alkali-activated geopolymer is suitable for preparation of high-strength concrete. Production cost is need to be paid more attention in the optimization process. The alkali-activated slag-fly ash geopolymer concretes with production cost reduction of 7.6%∼10.6% and carbon emission reduction of 77.3%∼80.7% at strength grade of C30, and with production cost reduction of 22.5%∼27.0% and carbon emission reduction of 76.9%∼81.3% at strength grade of C50 are achieved.