Grinding is a complex machining process with a lot of interactive parameters, which depend upon the grinding type and requirements of products. Surface roughness is a widely used index of product quality and in most cases it is considered as a technical requirement for mechanical product. In this paper, a neural network and fuzzy-based methodology are developed for predicting the surface roughness in a grinding process for work rolls used in cold rolling. This methodology predicts the most likely estimates of surface roughness along with lower and upper estimates using fuzzy numbers. The network model is trained using back-propagation algorithms. The initial training and testing dataset is identified on the basis of effect of a factor. The best possible network is chosen, with learning rate, number of neurons in hidden layer and error goals being decided automatically. A computer code was written in "C" language. The training and testing data are collected from carefully designed experiments. The validation of the methodology is carried out in a traverse cylindrical grinding machine in the Salem Steel Plant situated at Salem, India. The visual depiction of the training, testing and testing is presented. The validation of the methodology is carried out for grinding of alloy steel using a black carbide silicon grinding wheel. It is observed that the present methodology is able to make accurate predictions of surface roughness by utilizing small-sized training and testing datasets.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.