Neuro-fuzzy logic model for evaluating water content of sandy soils

In this study, the time-domain reflectometry (TDR) technique was utilized for generating an incident pulse signal of several harmonics of multiples of the fundamental frequency. The oscilloscope then captured reflected signals that were analyzed in the frequency domain by using Fourier transforms. The system response was characterized by its normalized spectral magnitude and phase angle. Experimental results indicated that the system is sensitive to water content and ion concentration. For water content prediction, a neuro-fuzzy logic model was developed. The changes in spectral magnitude and phase angle were considered as specific signatures for different soil conditions and were utilized in the training of the neuro-fuzzy logic model. The model was calibrated and used to predict other sets of data. The results indicated that the model is capable of predicting the water content of the tested soils.