Evaluation and Analysis of Particle Oxidation of HVOF Thermal Spraying Based on GA-BP Neural Network Algorithm

Abstract

In the process of High velocity oxygen fuel (HVOF) spraying, micron-sprayed particles are bound to oxidize under high temperature oxygen-containing environment, particle oxidation and burning are the key factors affecting coating quality. However, how to quantitatively evaluate and control particle oxidation is the bottleneck problem faced by the industry. In this paper, a three-dimensional transient calculation model of flame flow during HVOF thermal spraying WC-12Co process was established, and the computational fluid dynamics and discrete phase surface reaction model were combined to calculate and reveal the distribution characteristics of flame flow and the oxidation degree for particles during the spraying process. The calculation showed that the oxide layer thickness of particles varies greatly with different particle sizes. The oxide layer thickness of particles with 5 μm size is about 90 Å, and the oxide layer thickness of particles with 60 μm size is only about 8 Å. By adjusting the process parameters of oxygen/fuel ratio, particle size and nitrogen mass flow rate in the model, the output samples of sprayed particle flight temperature, velocity and oxide layer thickness can be obtained. On this basis, the sample data were statistically analyzed based on Genetic Algorithm-Back Propagation (GA-BP) neural network model, and the optimal process parameters for preparing the optimized coating were determined: particle size 27 μm, oxygen/fuel ratio 3.1, nitrogen mass flow rate 0.000363 kg/s. Experiments were carried out with optimized parameters, the results show that the optimized coating has fewer defects, lower oxide content and higher hardness and wear resistance. This study provides an important theoretical basis for quantitative preparation of high quality HVOF spray coatings.