A Study on Heuristic Knowledge Recording Expert System for WEDM. Investigation of System Configuration and Experimental Inspection.

Bibliographic Information

Other Title
  • WEDMにおける経験蓄積形エキスパートシステムの研究 システムの基本構成と実験的検証
  • Investigation of System Configuration and Experimental Inspection
  • システムの基本構成と実験的検証

Search this article

Abstract

In this paper, the authors propose a new configuration of an expert system for WEDM which realizes the heuristic knowledge recording by the machine itself. This system consists of the following 3 functions: (1) self-search of suitable setting of multi-parameters in machining, (2) fuzzy control for avoiding the unstable machining states and (3) database for recording and self-organizing the heuristic knowledge. According to the desirable performance of the machine directed by human operator such as “faster cutting speed”, the system progresses searching the most suitable setting of machining parameters. When the machining status falls into unstable during the search operation, fuzzy controller works to recover the machining stability. Reached values of the machining parameters obtained through the search operation are recorded and organized by the database. Fundamental experiments on searching the fastest cutting conditions were carried out to confirm the ability of the proposed system. As a result, the following conclusions were obtained. (1) Self-seach of the suitable values of the machining parameters to achieve the fastest cutting speed was carried out successfully. (2) The stability of the machining could be recovered quickly by the fuzzy controller during the search operation. (3) By organizing the reached values of the machining parameters obtained by the search operations, values of machining parameters close to the suitable ones could by estimated at the beginning of the search operation.

Journal

References(9)*help

See more

Details 詳細情報について

  • CRID
    1390282681212127104
  • NII Article ID
    10002439408
  • NII Book ID
    AN00151681
  • DOI
    10.2526/jseme.29.61_1
  • ISSN
    18810888
    0387754X
  • Text Lang
    ja
  • Data Source
    • JaLC
    • Crossref
    • CiNii Articles
  • Abstract License Flag
    Disallowed

Report a problem

Back to top