The Prediction Of Multiple Polishing Parameters Optimization By Using MANFIS-PID

Mazlan, Muhammad Ikhwan Haqiem (2020) The Prediction Of Multiple Polishing Parameters Optimization By Using MANFIS-PID. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) The Prediction Of Multiple Polishing Parameters Optimization By Using MANFIS-PID.pdf - Submitted Version Download (5MB)
	Text (Full text) The Prediction Of Multiple Polishing Parameters Optimization By Using MANFIS-PID.pdf - Submitted Version Restricted to Registered users only Download (23MB)

Abstract

The polishing is a secondary process that function as to give a uniform surface rouglmess of the product. Although the automation level is increasing gradually over many years, the industry still supports manual or half-automatic polishing method due to expensive cost and time consume. Moreover, in polishing process involves several parameters such polishing force and abrasive values to perform a high quality surface roughness of the product. Reported that, in artificial intelligent (Al) have its own limitation in performing prediction. In this project attempt to present a prediction for multiple parameter polishing optimization which involves polishing force (PF), polishing time (PT) and grid size (GS) as the outputs of system with input data given of Surface roughness (Ra) values. The main goal of this project to achieve high accuracy prediction of the MANFIS-PID compared to MANFIS subsystem itself. The MANFIS-PID model is proposed as it able to concentrate on the Single Input Multiple Output (SIMO) identification to make the prediction on the multiple polishing parameters with additional of PID control for reduce the error output of MANFIS subsystem without increasing complexity of system. This is system undergo two type analysis data which are training data, for determine reliability and testing for validation purpose by using MATLAB/Simulink software. As the result MANFIS-PID perform lower RMSE value compare to MANFISsubsystem itself.

Actions (login required)

View Item