Optimization of design parameters to evaluate the performance of journal bearing: ANN approach

Abstract

This investigation focus on the influence of process parameters on the performance of journal bearing. The process parameters and their range of values such as variance ratio (0–1), journal radius (25–65), radial clearance (30–60), dynamic viscosity (0.026–0.046), surface pattern parameters (1/3, 1/6, 1/9, 1, 3, 6, 9) and surface roughness parameter (15–30) were used in the evaluation. The artificial neural networks (ANN) approach was adopted for the evaluation through these process parameters as inputs and minimum film thickness, frictional torque at the surface and journal bearing mass as their target parameters. The method of training the program and concept of assigning the weights for each process parameters has been carried out for the journal bearing performance. The optimization model was designed and mean squared error (MSE) was used as the performance evaluation function. The values computed from the experimentation and predicted from the ANN model were compared. The suitability of Artificial Neural Network (ANN) technique was used in the process parameter optimization. Using ANN technique, the maximization of minimum oil film thickness, critical journal mass and minimization of frictional torque have been separately carried out for three bearings having transverse, isotropic or longitudinal type roughness patterns. The results revealed that the achievement of regression value above 99% and close agreement between expected and predicted value of all the three output functions shows the suitability of ANN in the design optimization of Journal bearing.