Evaluation Of Strength Properties Of Geopolymer Concrete By Using Artificial Intelligence And Machine Learning Techniques

Abstract

Geopolymer concrete (GPC) is a practical option in contrast to traditional cement, utilizing fly debris (FA) rather than customary Portland concrete (OPC), offering ecological and sturdiness benefits. This study utilized two AI (ML) strategies, quality articulation programming (GEP) and multi-articulation programming (MEP), to foster expectation models for the compressive and split rigidity of GPC with FA as a folio. An information base with 301 compressive strength and 96 split rigidity results was ordered. Seven information factors were utilized: FA, sodium hydroxide, sodium silicate, water, superplasticizer, and fine and coarse totals. Model execution was assessed utilizing measurable measurements and outright mistake plots. GEP-based models beat MEP-based models in execution, precision, and speculation. GEP models had higher connection coefficients (R) for compressive and split elastic qualities (0.89 and 0.87) contrasted with MEP models (0.76 and 0.73). Mean outright blunders for GEP models were 5.09 MPa (compressive) and 0.42 MPa (elastic), while MEP models had mistakes of 6.78 MPa and 0.51 MPa. The last models gave straightforward numerical plans utilizing GEP and Python code from MEP, showing potential for streamlining geopolymer blend plans. This examination features the significance of feasible materials and advances ML applications in the development business