Abstract: An approach of rough-fuzzy control, which uses rough set data analysis to upgrade fuzzy PID control, is presented for the electro-hydraulic servo position system in pipe bundle shaping process. The input to the model is mapped into the output subspace by using rules acquired from RSDA, then the output of the system is approximated in the subspace. On the basis of RSDA, modeling and analysis of actual pipe bundle shaping control process is provided. Compared with conventional fuzzy PID control, this method takes on better dynamic running performance with simple control arithmetic.

Key words: Electro-hydraulic bundle shaping, Rough set data analysis, Self-learning rough-fuzzy PID control