Prediction of compressive strength of roller compacted concrete using regression analysis and artificial neural networks

Abstract

The compressive strength is most reliable parameter to evaluate the ability of concrete in resisting compression. The paper presents a study on prediction of the compressive strength of roller compacted concrete using multiple regression analysis (MRA) and artificial neural networks (ANN). The compressive strength of roller compacted concrete was obtained experimentally at 3, 7 and 28 days of curing. The samples were prepared by varying the percentage of cement and superplasticizer. The data were organized in three different groups randomly using R statistical software. The models were executed with cement content, coarse and fine aggregate, superplasticizer content, water content and days of aging as input parameters that were used to predict compressive strength which is the output parameter. The analysis was performed using multiple regression and artificial neural networks methodology. Statistical measures like root-mean-square error (RMSE), mean absolute error (MAE) and coefficient of determination are used to assess the performance of models. The determination coefficient from multiple regression analysis is found to be 0.975 and 0.886 for testing and validating the data correspondingly, whereas the determination coefficient from artificial neural network analysis is found to be 0.9 for both testing and validating the data. The results obtained from ANN are highly accurate because of its own topology.

AI Mix Design of Fly Ash Admixed Concrete Based on Mechanical and Environmental Impact Considerations

“…In Figure 3 (also see Table 1), the compressive strength at 28 days curing and the environmental impact index (P) behaviour of FA admixed concrete corresponding to fly ash and binder ratio (FA/B; B = C+FA), fine aggregate and binder ratio (FAg/B) and coarse aggregate and binder ratio (CAg/B) is presented (see Figures 3-a to 3-f, respectively). It can be observed that the concrete mixes with high amount of cement produce high strength but with a compromised environment with high environmental impact (P) compared with the mixes with high FA and reduced cement, which produce low concrete strength (Fc28) but a more eco-friendly and eco-efficient environment with low impact (P) from the production and utilization of the concrete, which agrees with the studies [28][29][30][31][32][41][42][43][44][45][46][47][48][49][50][51][52][53], which studied the sorbent effect of increased FA in concrete. By implication, this shows that high FA/B ratio produce low strength compared to low FA/B as presented in Figures 3-a and 3-b.…”

“…An extensive literature search (experimental data, data arrangement, and tabulation), which has been conducted in this research work to collect the amount of concrete mix materials corresponding to the compressive strength at 28 days (Fc28) and the environmental impact, which has been evaluated by the method of life cycle assessment (LCA), has revealed the influence of fly ash (FA) proportion by partial or total replacement in the production of a more sustainable and eco-friendly concrete. During this first phase of the present research work, previous relevant research papers were reviewed [41][42][43][44][45][46][47][48][49][50][51][52] and data points relevant to this modeling and optimization exercise were collected and tabulated. Figure 1 shows the theoretical framework of the activities of this research work.…”