Modeling Magneto Rheological Damper Using Artificial Neural Network Method
Author | : Farah Wahida Yusof |
Publisher | : |
Total Pages | : 46 |
Release | : 2012 |
ISBN-10 | : OCLC:953546560 |
ISBN-13 | : |
Rating | : 4/5 (60 Downloads) |
Book excerpt: The purpose of this study is to model magneto rheological (MR) damper using the artificial neural network (ANN) method. There are two ways of modeling the MR damper, but throughout this study the non-parametric model is being use. The method chosen is ANN method. The model of MR damper was simulated in Simulink® MATLAB. Firstly, the block diagram to simulate the MR damper was designed. Instead of designing the block diagram, the ANN coding was also being designed. Next, the simulation was run to obtain the damping force predicted by ANN method. The values of learning rates and weights were being tuned to obtain the desired result from ANN method. The result from ANN method was compared with the result collected from the actual experiment. As it is a model, so the exact value of damping force is impossible to obtain. The difference between the actual damping force and the damping force obtain by ANN method will be calculated by plotted the error. RMSE was used to visualize the error. However, at the end of this project period, the smallest value of RMSE was 3.320945. This RMSE value still large if compared to other previous research on ANN method use to model MR damper. The best model should have the minimum RMSE value around 0.018.