Study on green concrete replacing natural fine and coarse aggregate by plastic waste – An experimental and machine learning approach

Abstract

One of the most universally used materials in construction industries is concrete because of its accessibility, affordability, extended lifespan and resistance to damage against unfavourable weather conditions. However, there is a massive depletion of valuable natural resources for the concrete production. Further, plastic consumption as well as waste generation is increasing day by day which is very much challenging to be managed as it is non-biodegradable and creates various environmental pollution. Thus, the usage of plastic waste as a fractional substitution of natural aggregates in construction industry may mitigate the problems. For the present study, nine concrete mixtures of M30 grade concrete containing recycled plastic waste polyethylene (RPE) and plastic waste polyethylene-terephthalate (WPET) were prepared as a fractional volumetric replacement for both natural coarse aggregates (NCA) and fine aggregates (NFA). The replacement levels were 0%, 5 %, 10%, 15%, 20%, 25%, 30%, 35% and 40% keeping constant the water-cement ratio (0.37).The usage of waste plastic aggregates for preparation of concrete were assessed through different fresh and hardened properties like slump, dry density, compressive strength (CS), flexural strength (FS), split tensile strength (STS) and ultrasonic pulse velocity. The test results corroborates that the slump increased with higher percentage replacement of waste plastic aggregate (PA) in concrete. Further, incorporation of WPET and RPE in concrete reduced the dry density by 10.59 % at 40% mix proportion. It is to be mentioned that, the CS decreases with with increase of partial replacement of natural aggregate with waste PA. The ultrasonic pulse velocity (UPV) also decreased with increase of the PA. The test result also depicted that up to 5% replacement the STS value is almost constant, after that it reduces. Further, in case of FS test, the replacement with waste PA concrete by 10%, the FS improved to 19.91%. Furthermore, the data prediction models of slump, CS, FS and STS were developed by artificial neural network. Moreover, the optimization of mix proportion was conducted by factor response optimizer with composite desirability of 0.65.