Abstract
Grinding, a complex surface finishing process, is used to generate a smooth surface on the job using a wheel. Several analytical and empirical expressions had been developed by various investigators to determine input-output relationships in grinding, which are highly non-linear, and those relationships may not be accurate. In the present work, power requirement and surface finish in grinding are expressed as the functions of three inputs, namely RPM of the job, RPM of the wheel and feed rate, and this process is modeled using a combined GA-Fuzzy approach. An optimized knowledge base (KB) of the fuzzy logic controller (FLC), which is the representative of the KB of a grinding machine, is obtained off-line, using a genetic algorithm (GA). The training data for the FLC is obtained from the empirical expressions (which are generally inaccurate) and the performance of the trained FLC is tested by comparing its results with those obtained through real experiments.
Get full access to this article
View all access options for this article.