Product Code: ICAL07_P510

Predictive Model for Thick Steel Laser Cutting Quality using Artificial Neural Networks
Marimuthu Sundar, Jadavpur University; Kolkata India
Asis Kumar Nath, RRCAT; Indore India
C.H. Premsingh, RRCAT; Indore India
Dipak Kumar Bandyopadhyay, Jadavpur University; Kolkata India
Dipten Misra, Jadavpur University; Kolkata India
Prabir Kumar Dey, Jadavpur University; Kolkata India
Shankar Prasat chaudhuri, Jadavpur University; Indore India
Presented at ICALEO 2007

Laser assisted oxygen (LASOX) cutting is emerging as a promising technique for cutting thick steel in the industries. The end quality of the cut steel, viz., the heat affected zone (HAZ), kerf and roughness depends on careful selection of the process parameters: speed, gas pressure, stand-off and laser power. In this paper, the combined effects of speed, gas pressure, stand-off and laser power on the quality characteristics of LASOX cutting parameters of thick carbon steel have been studied in detail for prediction of HAZ, kerf width and surface roughness. Cutting experiments have been performed based on 40 mm thick carbon steel with factorial experimental design technique using an indigenously developed CO2 laser machine. Initial studies reveal that it is possible to cut steel more than 50 mm using this process yielding good surface finish, lower HAZ width, and higher speed than the existing thick steel cutting processes like oxy-acetylene or plasma cutting. A model for prediction of cutting quality is developed using feed forward artificial neural networks exploiting experimental data. Output of the predictive model can be utilized for selection of optimized process parameters through an appropriate optimization technique.

Product Thumbnail


Members: $28.00

Note: When applicable, multiple quantity discounts are applied once the items are added to your cart.