UMĚLÁ INTELIGENCE VYUŽITÁ PŘI DEFINICE MATERIÁLOVÝCH VSTUPŮ, KTERÉ PŘEDURČUJÍ VÝSLEDEK ŘEZNÉHO POVRCHU PO AWJ ŘEZNÉM PROCESU

ARTIFICIAL INTELLIGENCE IN DEFINITION OF MATERIAL ENTER DATA THAT DETERMINE QUALITY FINISH AFTER AWJ CUTTING PROCESS

článek ve sborníku ve WoS nebo Scopus

en

Tolerance of setting AWJ cutting parameters on each material have big influence on quality finish and all characteristics of material kefr after cutting process. There is direct relation between species of workpiece material and final results obtained after AWJ cutting. Thus there is a need to have a tool for right material setting, choosing optimal cutting parameters and finding relations between each variables of cutting process. Using of information technology specially machine learning methods should be the right way. From all methods the Receptive Field Weighted Regression (RFWR) can be used as a function approximator for different mapping tasks like learning the value function for reinforcement learning. After learning process we can do several operations with obtained data base, like a choosing of cutting conditions for new or incomplete deffined material or finding relations between each properties of AWJ process.

V AWJ technologii je velké množství proměnných a parametrů, které mají vliv na ýsledek řezného procesu. Proto jsou potřebné zkušenosti získané z praxe a experimentů, uložit je v databázi, podrobit zkoumání a vytvořit na jejich základě znalostní systém, který by umožňoval podporu procesu volby vhodných řezných podmínek.

Tolerance of setting AWJ cutting parameters on each material have big influence on quality finish and all characteristics of material kefr after cutting process. There is direct relation between species of workpiece material and final results obtained after AWJ cutting. Thus there is a need to have a tool for right material setting, choosing optimal cutting parameters and finding relations between each variables of cutting process. Using of information technology specially machine learning methods should be the right way. From all methods the Receptive Field Weighted Regression (RFWR) can be used as a function approximator for different mapping tasks like learning the value function for reinforcement learning. After learning process we can do several operations with obtained data base, like a choosing of cutting conditions for new or incomplete deffined material or finding relations between each properties of AWJ process.

AWJ, řezné podmínky, umělá inteligence, znalostní systémy, expertní systémy, znalostní báze, databáze.

abrasive waterjet cutting, artificial intelligence, cutting conditons, knowledge systems, expert system, knowledge base, database.

2007

24.10.2007

Published by DAAM International

Vídeň

3-901509-58-5

Annals of DAAAM for 2007 & Proceedings of the 18th International DAAAM Symposium in Zadar

1

263–264

2