TECHNOLOGY FOR ENHANCING THE EFFICIENCY OF MANAGING WELDING OPERATIONS USING SOFT COMPUTING

The article focuses on developing a technology for enhancing the efficiency of managing welding operations based on the comprehensive use of swarm intelligence algorithms, evolutionary modelling, and applying convolutional neuro-fuzzy networks (CNFN). The relevance of the objective is associated with the lack of continuous control and analysis of welders’ work, the complexity of objectively assessing the quality of welds, the presence of a large number of weld-ing training courses, from which it is difficult to select those that meet production needs, and the necessity of pro-cessing a large amount of data for the effective distribution of welding tasks among performers. The authors develop a new technology for enhancing the efficiency of managing welding operations, based on using swarm intelligence algo-rithms for training welders, evolutionary modelling for distributing welding tasks among performers, and applying CNFN for recognizing weld defects. The proposed information technology provides comprehensive automation of the following stages of managing welding operations: forming the structure and composition of welders’ knowledge, skills, and abilities required by production; forming the composition of courses that ensure welders’ effective training in works demanded by production using bee swarm algorithms; training welders in works demanded by production in the shortest possible time using ant colony and firefly algorithms; rational distribution of tasks among welders based on evolutionary modelling; quality control of work through automatic recognition of weld defects using CNFN.