Probability-based unified sensitivity analysis for multi-objective performances of composite laminates: A surrogate-assisted approach

Abstract

The present paper proposes a surrogate-assisted moment-independent stochastic sensitivity analysis of laminated composite plates for establishing a unified measure in the case of multi-objective performances. With the advancements in artificially engineered structural systems spanning across different length scales, it has become more common to design composite structures for multi-objective performances like the criteria of deflection, buckling and vibration of multiple modes, different impact parameters, and failure. Normally sensitivity analysis is carried out separately and individually for different such performance parameters. This paradigm is no more suitable for advanced multi-functional structures like composite laminates. In this article, we propose an efficient unified sensitivity analysis approach based on weighted relative importance of different performance parameters by introducing the notion of engineering judgment. A moment-independent sensitivity analysis is proposed here based on finite element modeling of composites in conjunction with the Least Angle Regression assisted Polynomial Chaos Expansion (PCE) to achieve computational efficiency without compromising the outcome. Such surrogate-assisted finite element approaches are particularly crucial for computationally intensive multi-objective systems like composites. The layer-wise unified sensitivity quantification of laminated composites considering multi-functional objectives, as presented here, would lead to more optimized designs and better quality control while manufacturing the complex advanced structural systems.