Wyniki wyszukiwania - Biblioteka Nauki

Nowa wersja platformy, zawierająca wyłącznie zasoby pełnotekstowe, jest już dostępna.
Przejdź na https://bibliotekanauki.pl

Ograniczanie wyników

1 Prace Naukowe Instytutu Górnictwa Politechniki Wrocławskiej. Konferencje

1 1999

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Tool wear monitoring of machine tools

100%

Pan F. , Hope A. D. , King G. A.

Prace Naukowe Instytutu Górnictwa Politechniki Wrocławskiej. Konferencje

1999

tom Vol. 86, nr 24

124-132

The project team at Southampton institute have successfully completed the first phase of a programme of work, funded by the Engineering and Physical Science Research Council an the Department of Trade and Industry, to research and develop a novel cutting tool wear monitor. During this work a number of sensor were used to monitor physical parameters associated with tool wear during actual machining conditions and signal processing was carried out on the sampled data in order to extract typical features which can be used to represent tool wear. These features were then used to classify the level of tool wear using fuzzy pattern recognition techniques based on neural network and/or fuzzy logic. The work has concentrated on both turning and milling operations and prototype, demonstrator systems are available. An intelligent tool condition monitoring system is introduced in this paper. The system includes four kinds of sensor, together with signal conditioning equipment, interfaced to a computer. The following research issues are discussed in detail. (i). The selection of suitable sensors and data acquisition system to collect healthy signals from the machining process. (ii). The development and section of appropriate signal features that are sensitive to tool wear and insensitive to experimental noise. (iii). The development of signal classification procedures for reliable identification of tool wear states. A unigue neurofuzzy pattern recognition algorithm has been developed from this study. The algorithm can fuse the information collected from multiple sensor and possesses strong learning and noise suppression ability. This leads to successful classification of the tool wear state over a range of machining conditions.