Reliability Assessment of Sustainable Building Materials Using Fuzzy Probabilistic Models

Abstract

Because related data are seldom sufficient, expert judgments are imprecise, and forces and resistances differ, we provide an uncertainty-aware hierarchy that encompasses both random variation and incomplete knowledge. As a result of our work, we expect turning points, i.e., the points where acceptance or rejection will come about, to be bounded by risk very explicitly rather than simply giving a point estimate. The hybrid approach deals with both resistance and dependence through fuzzy-parameterised probability distributions, and reliability is assessed by the α-cut decomposition. Bands of failure probability and reliability index are computed using closed-form expressions wherever applicable but FORM/SORM or Monte Carlo methods otherwise. Dependence is taken care of by copulas, and we test the value of additional information through the average reduction in bandwidth. In a real-type case study of six concrete options, we find a consistent order across epistemic levels. Geopolymer remains the most reliable material; GGBS and flyash mixes lie in a close second tier, while OPC and recycled-aggregate combinations bring up the rear unless mean strengths are raised and dispersions tightened—higher expectations with negative dependence primarily contract the cautious band edge. The paper's unique contributions compared to previous crisp or type-1 treatments are distinguished by three factors: (i) α-level bands that integrate epistemic and aleatory effects; (ii) explicit treatment of dependence uncertainty; and (iii) an escalation ladder that is practical, can still be read by regulators and ties specific test/QA actions to risk reduction.