Modeling an artificial neural network to estimate cement consumption in clayey waste-cement mixtures based on curing temperature, mechanical strength, and resilient modulus

Seeking to address large-scale issues faced by many countries today, such as excessive energy consumption, global warming, and uncontrolled mining activities, this research repurposes clayey mining and excavation waste to design soil-cement mixtures for road construction. A total of 2026 data points from laboratory experimental tests were statistically analyzed using regression models and neural networks to evaluate the effect of curing temperature on compressive strength, indirect tensile strength, and resilient modulus. The study focused on three types of clayey waste mixed with high early-strength hydraulic cement (Type 1 Portland cement) after 7 days of curing. The samples were cured in three different chambers, each maintaining a constant temperature of 10, 28, and 40 °C for 7 days, simulating the most common road temperatures in Colombia. Results showed that temperature has a positive effect of 18 % on the resilient modulus, which could lead to cement savings in warm climates. Additionally, an artificial neural network model was developed, which can contribute to the construction and design of more sustainable and environmentally friendly geothermal pavements. The use of these models and networks not only facilitates the study of multiple variables but also optimizes materials and methods, aiming to reduce energy consumption and costs. • Triaxial resilient modulus clayey waste and Portland cement samples were made. • The curing temperatures influenced resilient modulus, compressive strength, and indirect tensile strength. • Samples cured at more than 40°C exhibited 18 % greater resistance when compared to 10°C. • A created artificial neural network showed a correlation between cement content, curing temperature, and compressive strength.