Chapter Texas Instruments

Concrete is the second most-used material in the world and with its production comes the need for better quality control. The strength of concrete is heavily influenced by its moisture content when it is formulated, giving rise to a need for an accurate measurement of that moisture content. This work focuses on the dielectric properties of water between 10MHz and 4 GHz to obtain information about the water content within a sample of wet concrete. A rectangular waveguide is used as a guided microwave spectrometry system to obtain data within this frequency range. In the 500 MHz to 1.5 GHz range the real part of the dielectric constant of water (ε’) is dominant and predictably shifts the cutoff frequency of the waveguide lower as water content increases. At frequencies increasing above cutoff the imaginary part of the dielectric constant of water (ε’’) increases as well meaning the concrete is more lossy. A principal component analysis is performed on this information and the modified data is run through a neural network to determine the water to cementitious material ratio of the concrete. Currently, alternative waveguide structures are being investigated and simulated using CST Microwave studio and data is being taken to better train the neural network.