PSI - Issue 68

Hussam Safieh et al. / Procedia Structural Integrity 68 (2025) 245 – 251 H. Safieh et al / Structural Integrity Procedia 00 (2025) 000–000

246

2

Durodola et al., (2018).; Pujol & Pinto, (2011); Salameh et al., (2024)) focuses on using cutting-edge prediction models to make sure that FRP-strengthened concrete structures are safe and durable in uncomfortable conditions. The model's performance was assessed by Safieh et al. (2024) using two trustworthy metrics: the mean squared error (MSE) and the coefficient of determination (R2). Fly ash percentage levels (FA%), superplasticizer content, water to binder ratio (w/b), and curing time are the factors influencing UHPC's compressive strength. Linear regression, Support Vector Machines (SVM), Neural Networks, and Random Forest techniques comprised the 54 ML models that were employed. This work also employs the Random Forest ML model to perform a parametric study that will assist in determining the compressive strength of UHPC with a greater FA% content. Again, a thorough parametric analysis was carried out by Abuodeh et al. (2020) and the results showed that RBPNN (RBPNN) with RFE (RFE) and NID (NID), individually, is a useful tool for evaluating the behavior and strength of FRP in shear-strengthened beams. Two novel, highly dimensional, dynamic, and difficult functions that might be used for evaluating new algorithms are proposed by Naser et al., (2024) after reviewing the top 25 functions that are often utilized in the public literature. Ultimately, this analysis highlights the shortcomings in the way benchmarking is currently done and offers ideas for further research. Ultimately, this analysis highlights the shortcomings in the way benchmarking is currently done and offers ideas for further research. It was suggested by Mhanna et al., (2023) to augment the proportion of GGBS to fly ash in the GPC mixture and incorporate a high-efficiency superplasticizer to preserve the workability of the mortar without compromising its compressive strength. Furthermore, the general behavior of SCM beams was found to be similar to that of the control OPC beams, according to Hawileh et al., (2024). For RC beams cast with GGBS and fly ash, the flexural and shear capacities were also estimated using the ACI 318-19 design criteria. According to the aforementioned trials, cement substitutions in concrete with supplementary cementitious materials (SCM) are feasible and effective, which helps reduce carbon footprints. On the other hand, Shaw & Sil, (2022) examined the effects of fly ash on concrete in terms of the water-binder ratio, fly ash replacement percentages, and curing ages (days), both in the mixed form and with and without admixture. Statistical best fit functions were employed for the modelling and investigation of these features. This thorough investigation demonstrates that the compressive strength of all varieties of concrete decreases as the water-binder ratio rises. In contrast, maximum strength gaining commences between 7 and 90 days in the case of "with admixture fly ash concrete," while significant strength gaining is observed after 90 days and ultimate strength gaining occurs at 365 days. In the maximum cases, the highest compressive strength observed is at 20% fly ash (FA) replacement level. Based on the outcomes from this paper, some research works have been carried out considering the cyclic loading performances on the exterior beam-column joints (Shaw et al., (2022), (2023), (2020)). In this investigation, based on the data of research paper (Shaw & Sil, (2022)), a machine learning and artificial neural networks models have been approached to predict the compressive strength for the FA-OPC concrete (fly ash with ordinary Portland cement) in aspects of various factors such as fly ash replacement percentages, water-binder ratio and curing periods for both i.e. with and without superplasticizer FA-OPC concrete. 2. Methodology The structured approach for predicting the compressive strength of NWC begins with the input parameters, which are the replacement percentage of fly ash (FA%), superplasticizer content, w/b ratio, and curing period. These variables are used to predict the output variable, compressive strength. The data is then divided, with 80% allocated for training the ML model—Linear Regression, Linear SVM and Random Forest—and the remaining 20% are set aside for testing. Following the training phase, the models are subjected to validation processes to determine the optimal model, using metrics such as the ! value depicted in the embedded graph within the flowchart. This comprehensive method ensures the selection of a model that is best suited to accurately predict the NWC’s compressive strength based on the given inputs. 2.1. Dataset Collection The dataset used in this study was sourced from experiments conducted on OPC mixes with fly ash as a supplementary material. It includes the following variables: the percentage of fly ash (FA%), superplasticizer content, water-to-binder (w/b) ratio, curing period (in days), and the resulting compressive strength (in MPa). Fly ash