An AI Approach to Tasking and Control of an Industrial Laser System

Lim, See Yew (1993) An AI Approach to Tasking and Control of an Industrial Laser System. PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 11007779.pdf] PDF
Download (7MB)


Effective use of lasers for materials processing applications requires a thorough knowledge of the fundamental mechanisms governing the interaction of radiation with matter. A discussion of the highly non-linear laser-material interaction is thus presented. The research then focusses on alternative approaches to the non-linear analysis of laser material processing through the application of genetic algorithms and chaos theory. A genetic algorithm is developed which predicts the laser cutting rate with good accuracy. The theory of deterministic chaos was exploited to investigate laser cutting of stainless steel; a new fundamental understanding of the interaction via the energy phase portrait was established. The construction of the phase portrait is illustrated. The work then focusses on the actual design and implementation of a hybrid intelligent system for the tasking and control of a gas laser for materials processing. Existing artificial intelligence (AI) methodologies are reviewed. A new AI topological technique is developed to facilitate the design of a Proportional-Integral- Differential Knowledge Based Management System (PID-KBMS). The final intelligent system comprises of the PID-KBMS and a Cone Decision Support System (CDSS). The CDSS provides on-line image acquisition and analysis of the laser cutting process. The architecture of the hybrid intelligent system is illustrated. The intelligent system source code was implemented in a Microsoft Windows and UNIX XWindows environment using Borland C. The intelligent system is hosted by a Viglen 486 PC and a SUN-SPARC 4. A virtual graphical user interface is presented on the Viglen 486 via the network. A final evaluation of the investigations into: the non-linear characteristics of laser-material interactions and prediction schemes, the AI topological technique for the design of AI control engines, the design and implementation of the PID-KBMS and CDSS are presented in the concluding chapter. The algorithms and methods applied herein have general applicability to any non-linear control problem. Finally, recommendations for future research are given.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Additional Information: Adviser: Chris R Chatwin
Keywords: Materials science, Optics, Industrial engineering
Date of Award: 1993
Depositing User: Enlighten Team
Unique ID: glathesis:1993-74859
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 27 Sep 2019 15:48
Last Modified: 27 Sep 2019 15:48

Actions (login required)

View Item View Item


Downloads per month over past year