Neuro-fuzzy system for tool condition monitoring in metal cutting Conference Paper uri icon

abstract

  • A neuro-fuzzy system is used to predict the condition of the tool in a milling process. The relationship between the sensor values and the condition of the tool, captured by the neural network is sought to be reflected in linguistic form using fuzzy logic. In this paper we present an error-based, density-driven adaptation scheme which tunes the fuzzy membership functions so that the rule set and linguistic labels reflect the true nature of the process variables in terms of their value and character. Experimental results show that when the technique is used, the fuzzy mechanism correctly predicts the condition of the tool in 97% of the cases presented to it.

author list (cited authors)

  • Mesina, O. S., & Langari, R.

publication date

  • December 1994