Characterization Of Stainless Steel 410 L Pta Hardfaced Valve Seat Rings Using Regression Analysis And Artificial Neural Network
The effect of wear is drastic paving way for fallouts on metallic components in its application and as well as in
their cost. In recent years, while unraveling wear problems, it is found that the hardfacing technique has substantially grown
up. In such a progress, Stainless steel 410 L is deposited on the face of low carbon steel ASTM-A105 valve seat rings by
plasma transferred arc welding process. The mathematical model for predicting the main and interaction effects of PTAW
variables for stainless steel hardfacng for which the dilution and bead geometry from the experimental data were obtained.
The experiments were based on the central composite rotatable design matrix of five factor, five level factorial technique.
Regression analysis was used to develop the models and the variance method was used to test their adequacy. The
percentage dilution was optimised (minimised) subject to the constraints of penetration, reinforcement, width and the total
area of the weld bead geometry. During optimization, enormous amount of data was generated from iterations and
substantial calculations needed with each design cycle requiring. The optimization module available in the toolbox of
Quality America six-sigma software suit DOE-PC IV version 3.01 was used. Mathematical models relating the wear and the
main factors, viz; Normal Load, Disc Speed, and Track Radius were developed. Optimization of the model was
accomplished to minimise the wear rate by using a three-level factorial technique. The outcome obtained show that the
mathematical model assesses the validity of the factors for a desired wear condition and the wear increases when the normal
load and disc speed increased.
Keywords- Wear, Hardfacing, Plasma transferred Arc Welding, Artificial Neural Network, Regression Analysis, RSM