This study explores the possibility of two matrices containing metallic particulates to act as smart materials by sensing of strain due to the presence of the conducting particles in the matrix. The first matrix is a regular Portland cement-based one while the second is a novel iron-based, carbonated binder developed at ASU. Four different iron replacement percentages by volume (10%, 20%, 30% and 40%) in a Portland cement matrix were selected, whereas the best performing iron carbonate matrix developed was used. Electrical impedance spectroscopy was used to obtain the characteristic Nyquist plot before and after application of flexural load. Electrical circuit models were used to extract the changes in electrical properties under application of load. Strain sensing behavior was evaluated with respect to application of different stress levels and varying replacement levels of the inclusion. A similar approach was used to study the strain sensing capabilities of novel iron carbonate binder. It was observed that the strain sensing efficiency increased with increasing iron percentage and the resistivity increased with increase in load (or applied stress) for both the matrices. It is also found that the iron carbonate binder is more efficient in strain sensing as it had a higher gage factor when compared to the OPC matrix containing metallic inclusions.
Analytical equations (Maxwell) were used to extract frequency dependent electrical conductivity and permittivity of the cement paste (or the host matrix), interface, inclusion (iron) and voids to develop a generic electro-mechanical coupling model to for the strain sensing behavior. COMSOL Multiphysics 5.2a was used as finite element analysis software to develop the model. A MATLAB formulation was used to generate the microstructure with different volume fractions of inclusions. Material properties were assigned (the frequency dependent electrical parameters) and the coupled structural and electrical physics interface in COMSOL was used to model the strain sensing response. The experimental change in resistance matched well with the simulated values, indicating the applicability of the model to predict the strain sensing response of particulate composite systems.