Search

ANN can be an alternate to provide the general solution of erosion wear problems. In this case study, the artificial intelligence is utilized to predict the erosion wear in different slurries. The data was collected for the various thermal sprayed coatings namely Stellite, Colmonoy, Ni-20Cr2O3, Ni-20Al2O3, Al-20TiO2, WC-10Co4Cr, WC-10Co4Cr+Y2O3, WC-10Co4Cr+Mo2C, and WC-10Co4Cr+ZrO2. For the above mentioned coatings, three different erodent materials were used such as sand, fly ash, and bottom ash. These coatings materials were coated on SS316L steel using high-velocity oxy-fuel (HVOF) technique. MATLAB 16.0 was used in this study to design a standard multilayer feedforward hierarchical neural network to predict erosion wear. Pearson coefficient (R) values for the testing of the ANN model constructed for sand, fly ash and bottom ash slurries were 0.97534, 0.9638, and 0.95181 respectively. It was found that ANN model designed for sand slurry shows ±4.87% deviation from the experimental data. However, the percentage error discovered between experimental and ANN projected erosion wear was ±4.27 and ±4.98 percent respectively for fly ash and bottom ash slurry, which implies that the experimental data and neural network predictions agreed quite well. [ABSTRACT FROM AUTHOR]

Nanoparticles of MgxZn1-xFe2O4 ferrite (where x = 0.2, 0.4, 0.5,) are synthesized via citrate precursor method and then all the samples are sintered at 600 °C for 1 hour. The prepared samples are characterized through XRD (X-ray diffraction), FTIR (Fourier transform infrared spectroscopy), TEM (transmission electron microscope) and VSM (vibrating sample magnetometer). The M-H curves show evidence of a superparamagnetic (SPM) regime in the synthesized ferrites. [ABSTRACT FROM AUTHOR]