Compressive Strength Prediction for Industrial Waste-Based SCC Using Artificial Neural Network

Abstract

Concrete is the most used construction material in the world. Sustainable construction practice demands durable material. A particular type of concrete that flows and consolidates under its weight is proposed to reduce labor dependency during construction, called self-compacting concrete. It is installed without vibration due to its excellent deformability and flowability while remaining cohesive enough to be treated without difficulty. Evaluating its compressive strength is essential as it is used in important construction projects. An artificial neural network (ANN) is a predicting tool that can predict output in various sectors. This study evaluated the compressive strength of industrial waste such as fly ash and silica fume incorporated in self-compacting concrete at various ages. A non-linear relationship was used to develop the model relating mix composition and SCC compressive strength using an Artificial Neural Network (ANN). The experimental and expected outcomes were compared with the model prediction to evaluate the predictive capacity, generalize the generated model, and observe suitable matches. The developed ANN network can predict the desired output, i.e., compressive strength incorporating industrial waste. Furthermore, the influence of individual parameters viz. cement, silica fume, and fly ash, w/b were also evaluated using parametric analysis, which shows the sensitivity of various materials on the compressive strength of Self-compacting concrete. As a result, a higher correlation coefficient of 0.9835 with a smaller value of MAPE (0.0347) and RMSE (2.4503) is obtained. Finally, a process of creating tools for practical engineers and field users is proposed, which would be very handy and fast for predicting the strength of SCC.

Application of QUAL2Kw for Water Quality Modeling in the Cibarani Irrigation Channel, Bandung City, Indonesia

“…To evaluate the modeling results, metrics such as root mean square error (RMSE), mean absolute percentage error (MAPE), and coefficient of determination (R 2 ) are employed to assess the accuracy of the calculations [22]. Predicted values with smaller RMSE and MAPE usually indicate more accurate predictions, while higher R 2 values signify better-correlated results [23]. The expressions for each of these metrics are shown in equations ( 5) to (7):…”

Investigation of Deoxygenation Rate Determination in Cikakembang River, West Java, Indonesia

“…Root mean square error (RMSE), mean absolute percentage error (MAPE), and coefficient of deter-mination (R 2 ) evaluated the accuracy of calculations (Band et al, 2021). Predicted values with smaller RMSE and MAPE typically showed more accurate predictions, while higher R 2 values signified better-correlated results (Hossain,et.al, 2022).…”