A Probabilistic Model for Microstructure Evolution with Application to Smart Concrete Materials

Cornell University, Ithaca NY

Investigators

Abstract

The project proposes to formulate methods for representing and analyzing random heterogeneous materials in three directions. First, general probabilistic models are developed for the microstructure of cement based composites with carbon fibers. The models are random fields characterizing the geometrical, mechanical, and electrical properties of the material constituents. These cement based composites have an interesting microstructure, and are used in both mechanics and electronics as sensors of mechanical properties. Second, a computational environment is developed for calculating the evolution in time of the properties of microstructures subjected to deterministic loads. Finite element procedures are used to calculate changes in, for example, the stress, strain, damage and electric fields in cement based composite specimens generated from their probabilistic models. Development of the finite element code entails advances in numerical solver techniques as well as formulation of new mathematically sound traction-displacment relationships to model cohesive fracture. Third, the results of the finite element analyses are used to (i) update the probabilistic models of the material microstructure so they reflect the current damage state and (ii) develop macroscopic damage models for cement based composites by using homogenization procedures. The evolution of the material state is monitored by a mathematical object, referred to as digital material (DM), that defines uniquely the state of the microstructure at any time during loading and can be stored in the computer. The results of the analytical and numerical developments are validated against experimental tests providing information of the properties of the fiber matrix interaction and on the dependence of electric conductivity of macroscopic specimens on their damage state. The main outcome of the project is a novel methodology for predicting structural behavior of composite materials and for designing materials with better structural properties and better capabilities for sensing damage and strain. Other outcomes include new finite element codes, new mathematical models, and a strong educational outreach program.