Fuzzy evaluation of weld quality based on Matlab

Based on standards and the expertise, 10 indicators such as crack, reinforcement and splash, etc. are selected for welds quality description. The indicators are classified into three categories: appearance defects, shape defects and weld defects, and corresponding fuzzy evaluation sets are designed. Membership functions of each indicator are determined with fuzzy inference system (FIS) editor in Matlab based on the actual welding experience. A two-stage fuzzy evaluation model for weld quality is established in Simulink. Finally, the model is tested through evaluating 7 different types of welds. The results show that fuzzy evaluation calculation can be simplified by combining Simulink with Matlab-FIS, and the evaluation results are more accurate and objective compared with the experts' subjective evaluation. Copyright: © 2014 Editorial Board of CHINA WELDING.