Modelling of laser processing cut quality by an adaptive network-based fuzzy inference system

Abstract

Real-world problems in precision machining now require intelligent systems that integrate knowledge, techniques, and methodologies. Intelligent systems possess human-like expertise within a specific domain to adapt themselves and to learn to do better in making decisions for an intelligent manufacturing system. An intelligent tool called adaptive network-based fuzzy inference system (ANFIS) was used to model and predict the laser cut quality of a 2.5 mm manganese—molybdenum (Mn—Mo) alloy pressure vessel plate in this article. A 3 kW CO₂ laser machine with seven selected design parameters was used to carry out 128 experiments based on 2 ^k factorial design with single replication. Because surface roughness (Ra) was the response parameter, it was targeted to be <15 μm to meet the requirement and benchmark of the pressure vessel manufacturer who sponsored this project. The DIN 2310-5 German laser cutting of metallic materials standard and procedure was referred to for evaluating surface roughness, where experimentally obtained results were used for Ra predictive modelling. Predictions of non-linear laser processing by ANFIS were found to be extremely promising in supplying the desired output, where Ra was predicted to an excellent degree of accuracy, reaching almost 70 per cent with the experimental pure error below 30 per cent.