Title: Analysis and application of the adaptive neuro-fuzzy inference system in prediction of CMP machining parameters
Authors: Z.-C. Lin, C.-Y. Liu
Addresses: Department of Mechanical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, Republic of China. Department of Mechanical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan, Republic of China
Abstract: This paper examines the machining parameters during a wafer flatting process with chemical mechanical polishing (CMP). This study adopted the adaptive neuro-fuzzy inference system (ANFIS) to reduce the number of CMP experiments and to predict the surface roughness in the absence of CMP experiments. An integrated concept like ANFIS combines the advantages of the two systems of fuzzy control and neural networks. Using this method, we can apply a neural network with learning and computation capability in fuzzy control. On the other hand, fuzzy control also provides high-level expert knowledge and fuzzy rules for use in the neuro network. The findings indicate that after entering the experimental values in ANFIS for learning, the surface roughness value in the absence of experiments and even the relative values of the other process parameters can be predicted. Besides, the error between the predicted values and the experimental values is within 3%. Verification through experiments indicates that the experimental values of the process parameters match the results from the inference of the ANFIS. Thus, the experimental data of CMP process parameters can be precisely controlled in this study.
Keywords: CMP; neuro-fuzzy; optimisation.
International Journal of Computer Applications in Technology, 2003 Vol.17 No.2, pp.80-89
Published online: 16 Jul 2003 *Full-text access for editors Access for subscribers Purchase this article Comment on this article