Compensation of systematic errors in five-axis high-speed machining
2002 (English)In: International Journal of Production Research, ISSN 0020-7543, Vol. 40, no 15 SPEC., 3765-3778 p.Article in journal (Refereed) Published
A method is described for the compensation of errors associated with tool path generation, particularly during five-axis high-speed machining (HSM). Information on machine tool performance and its dynamic features is used to calculate possible errors and convenient modifications of the NC program, thereby avoiding errors when parts are actually being machined. This 'preprocess method' by means of postprocessing with NC software is presented. The errors dealt with are mainly servo lag errors, but the explored approach can support most systematic errors associated with machine tool performance. These are briefly summarized. Research so far has largely been aimed at the implementation of compensation routines in the CNC controller and at corrections in real time. The problem is that most applications are only available for three-axis milling. The presented approach (compensation before machining by using software routines in a specially designed postprocessor) is based on a fuzzy logic expert system. The benefits can be summarized as data reduction, data improvement, precontrol of feed, and improved component accuracy. The applied procedure is also convenient for implementation in an industrial environment by retrofitting existing equipment. The suggested method provides improved control over machine dynamics, permitting high-speed machining centres to maintain a maximum or near-maximum feed rate despite axis reversals and tool path changes, even at corners. Two categories of results are presented, namely the management of NC data related to expected performance of the machine tool and improvements of the machine tool performance in terms of productivity and accuracy of the machined test components.
Place, publisher, year, edition, pages
2002. Vol. 40, no 15 SPEC., 3765-3778 p.
Engineering and Technology
IdentifiersURN: urn:nbn:se:liu:diva-46893DOI: 10.1080/00207540210136469OAI: oai:DiVA.org:liu-46893DiVA: diva2:267789