Review Article
Blending Data-Driven Surrogates with Physics - Based Topology Optimization: A Critical Review of Machine Learning - Accelerated Design in Fibre - Reinforced Polymer and Concrete Structures
Girmay Mengesha Azanaw*
Issue:
Volume 10, Issue 3, September 2025
Pages:
80-93
Received:
10 May 2025
Accepted:
29 May 2025
Published:
28 July 2025
Abstract: Blending data driven surrogates with physics based topology optimization holds the promise of revolutionizing the design of fibre reinforced polymer (FRP) composites and concrete structures by dramatically reducing computational cost while preserving-or even enhancing-solution quality. This critical review synthesizes developments from last decade in which machine learning (ML) models such as deep neural networks, Gaussian processes, and ensemble learners have been trained to approximate finite element analyses within iterative optimization loops. The author investigates the applications of Fiber Reinforced Polymer (FRP) composites, wherein the exigencies of continuous fiber orientation and constraints imposed by additive manufacturing necessitate the employment of high-fidelity yet efficient computational solvers. Additionally, The author explore the domain of concrete structures, wherein the inherent heterogeneity, prevalence of cracking, and considerations of durability present distinctive challenges for modeling. By conducting a comprehensive literature review utilizing databases such as Scopus, Web of Science, IEEE Xplore, and MDPI, alongside stringent inclusion criteria, we extract and analyze algorithmic selections, training data configurations, performance metrics (including prediction error and speed-up factors), and outcomes pertaining to manufacturability. The findings indicate that workflows driven by neural surrogate models can achieve accelerations of up to 50 times while maintaining deviations of less than 5% from full-order models; however, limitations in generalizability across various loading scenarios persist. The author delineate critical deficiencies, including the scarcity of benchmark datasets, restricted transfer learning across diverse material systems, and integration challenges with Computer-Aided Design (CAD) and Finite Element Analysis (FEA) frameworks, and The author propose avenues for future research which encompass hybrid physics-based machine learning frameworks and real-time optimization. By elucidating best practices as well as existing challenges, this review offers a strategic roadmap for researchers and practitioners aiming to exploit machine learning-accelerated topology optimization in the advancement of next-generation composite and concrete design.
Abstract: Blending data driven surrogates with physics based topology optimization holds the promise of revolutionizing the design of fibre reinforced polymer (FRP) composites and concrete structures by dramatically reducing computational cost while preserving-or even enhancing-solution quality. This critical review synthesizes developments from last decade ...
Show More
Research Article
Groundwater Quality Investigation in the Coastal Aquifer of Limbe, South West Cameroon
Ewanoge Mesumbe,
Alice Magha Mufur*,
Mathias Fru Fonteh
Issue:
Volume 10, Issue 3, September 2025
Pages:
94-109
Received:
25 March 2025
Accepted:
10 June 2025
Published:
30 July 2025
Abstract: Coastal aquifers are vital fresh water reservoirs that could be affected by seawater intrusion, thereby polluting the water resources. This study investigated the current status of subsurface water in Limbe-Cameroon, focusing on aquifer hydrochemical characteristics. Groundwater samples were obtained from nine boreholes and measurements were conducted on the following physicochemical parameters; pH, electrical conductivity (EC), and total dissolved solids (TDS) and major ions (cations and anions). The results showed that most of the sampled boreholes were in the permissible limits of the World Health Organization (WHO) guidelines, except for a few samples. 11.11% of the pH values, 11.11% of the EC values and 11.11% of the TDS values the WHO recommended limits. Major ion concentrations were below WHO prescribed levels in all analysed samples. The water quality index (WQI) indicated that 44.44% of the samples were of good quality water with water quality values varying from 26-50, 11.11% were classified as poor-quality water and another 11.11% of the samples were unsuitable for drinking purposes. The hydrochemical facies were principally Ca-HCO3 and Ca-Mg-Cl-SO4 water types. Irrigation water quality indices such as sodium adsorption ratio (SAR), Magnesium Hazard (MH), soluble sodium percentage (SSP) indicated that groundwater in Limbe is suitable for irrigation. These higher values signify the possiblity of salt water intrusion in the study area and highlights the critical need for sustainable groundwater management in Limbe to prevent further degradation from seawater intrusion and protect the freshwater resources in the region.
Abstract: Coastal aquifers are vital fresh water reservoirs that could be affected by seawater intrusion, thereby polluting the water resources. This study investigated the current status of subsurface water in Limbe-Cameroon, focusing on aquifer hydrochemical characteristics. Groundwater samples were obtained from nine boreholes and measurements were conduc...
Show More
