INTEGRATING FINITE ELEMENT ANALYSIS AND MACHINE LEARNING FOR THE DEVELOPMENT OF DESIGN INTERACTION CHARTS IN FRC TUNNEL SEGMENTAL LININGS

Abstract

Over the past two decades, fiber-reinforced concrete (FRC) has been used in a variety of tunneling projects, whether in conjunction with conventional rebars or without. The reason for adding fibers to cementitious composites is mainly to enhance their tensile behavior, with different fibers contributing differently to tensile strength and stress-strain behaviors. Despite FRC's increased use in tunneling projects, FRC segmental tunnel lining is still held back in its design conception due to a lack of applicable design procedures or charts for the selection of appropriate material strengths and thicknesses. This study aims to investigate and analytically determine the combined axial-moment strength capacities (M-N curve) of FRC unit sections and check said strength capacities against the expected applied forces from surrounding soil extracted using finite element analysis (FEA) used to model the complex soil-structure interaction and resulting deviatoric stresses. Using the analyses results, this novel study proposes new easy-to-use standardized design interaction charts based on comprehensive parametric analyses using multiple configurations of tunnel geometries and FRC material key parameters. These design charts will enable users to easily select with confidence the lining material and thickness, making them an indispensable tool for engineers and stakeholders involved in tunnel design and construction projects.