Polymers, Год журнала: 2025, Номер 17(8), С. 1046 - 1046

Опубликована: Апрель 12, 2025

The rheological properties of mortar are vital importance to ensure the quality and durability engineering structures, improving construction efficiency adapting different environments. This research focused on examining geopolymer (GM) with incorporation metakaolin (MK), nano-SiO2 (NS) polyvinyl alcohol (PVA) fibers. varying concentrations PVA fiber ranging from 0 1.2% (interval 0.2%) NS 2.5% 0.5%). As mix proportion optimization GM is normally carried out experimentally, a significant amount labor material resources was consumed. Based large amounts authentic operation data, prediction model for NS- PVA-fiber-reinforced developed using back propagation (BP) neural network. Subsequently, parameters were refined genetic algorithm (GA) predict reinforced dosages fiber. Three parameters, including static yield stress, plastic viscosity dynamic used evaluate GM. Moreover, Root Mean Square Error (RMSE), Absolute Percentage (MAPE) (MAE) applied assess capability algorithms. When GA–BP network used, compared BP network, coefficient determination (R2) viscosity, stress increased by 4.40%, 2.11% 15.28%, respectively, provided superior fitting effect, higher accuracy faster convergence. outputs model, can be adopted as precise method forecast

Язык: Английский