Life cycle assessment multi-objective optimization and deep belief network model for sustainable lightweight aggregate concrete

Abstract There has been rapidly growing demand for manufacturing structural lightweight concrete (LWC) owing to its reduced gravity and seismic loads. Concrete is required to meet increasingly more stringent eco-efficiency and sustainability requirements. Therefore, an experimental study on 30 LWC mixtures (S1 to S30) was conducted. The experimental variables included the percentage of lightweight expanded clay aggregate (LECA), water-to-cement ratio, cement content, and silica fume dosage. The effects of the test variables on the compressive and tensile strengths of LWC were assessed. The resulting experimental database was used to train and test deep belief network models for accurate prediction of the mechanical properties of LWC. Moreover, life cycle assessment based multi-objective optimization was deployed for optimizing sustainable LWC mixture proportions with minimized environmental footprint and life cycle cost, while maximizing the mechanical strength. The genetic algorithms approach was furtherly used to determine optimal LECA replacement levels for natural aggregates by minimizing the economic cost and environmental life cycle impact and maximizing the compressive and tensile strengths. The compressive and tensile strengths of LWC specimens ranged from 14.9 to 50.1 MPa and from 1.62 to 3.9 MPa, respectively, with the lowest and highest values belonging to mixtures S23 and S7, respectively. The LWC mixture S9 achieved the least environmental impact, while mixture S18 induced the highest environmental burden. The cost increased with increasing mechanical strength, life cycle, and LECA replacement level. Overall, the computational approach allows the user to accurately predict the mechanical strength of LWC and optimize sustainable LWC mixtures for specific target performance objectives, tailored economic cost, and eco-efficiency performance.