Prediction of Wear Behavior in Porous Sintered Steels: Artificial Neural Network Approach

Delicate parts with very high dimensional accuracy can be produced through the powder metallurgy process. Powder metallurgy in comparison with other manufacturing process, provide the possibility of access to precision pieces with a significant reduction in cost. Due to the increasing usage of powder metallurgy, there is a demand to evaluate and improve the mechanical properties of this parts. One of the most important mechanical properties is wear behavior, especially in parts that are in contact with each other. The wear of a solid surface due to mechanical contact between the two surfaces is one of the most important issue in many components. Therefore, the choice of materials and select manufacturing parameters are very important to achieve proper wear behavior. So, prediction of wear resistance is important in powder metallurgy parts. In this paper, we try to investigate and predict the wear resistance (volume loss) of powder metallurgy porous steels according to the affecting factors such as: density, force and sliding distance by artificial neural network. Artificial neural network training was done by multilayer perceptron procedure. The comparison of the results estimated by the artificial neural network with the experimental data obtained from the tests that normalized on the diagram, shows their proper matching. This issue confirms the efficiency of using method for prediction of wear resistance in powder metallurgy porous steel parts.

Keywords:

Wear Behavior , Powder Metallurgy Steels , Artificial Neural Network