Neural computing and Taguchi’s methodbased study on erosion of advanced Mo2C–WC10Co4Cr coating for the centrifugal pump

Abstract

Nowadays, computational and computing tools are widely used in prediction applications. Neural computing is a modern and emer ging technique to predict data efficiently and precisely. In this context, erosion investigation of Mo2C–WC10Co4Cr high-velocity oxy-fuel (HVOF) deposited on AISI 316L was carried out in the present work by implementing a neural computing and Taguchi’s method. This research paper focuses on neural network (NN) tech nique for the prediction of erosion in Mo2C–WC10Co4Cr HVOF coating. The WC10Co4Cr powder was composed of 3% (by weight) of molybdenum carbide, each with a concentration of 3 wt.%. The input parameters used for designing the NN model were erodent properties (bulk density, circularity factor, average particle size, and slurry concentration), material properties (hardness and porosity of bare/coated), and process parameters (speed and time). A set of experiments was optimised by using Taguchi’s method (L16 2 × 5 array). Results indicated that the overall accuracy of Pearson coeffi cient (R) was found as 0.99582 with experimental data. However, R = 1 was found for testing results