Simulation and Forecasting of Wire Tool Piles Fatigue Endurance by Using Neural Networks

• Published in: Procedia engineering 2017, Vol.206, p.1443-1451
• Main Authors: Rastorguev, D.A., Zotov, A.V., Bobrowskii, A.V.
• Format: Article
• Language: English
• Subjects: Cladding; Flexible tool; Neural network; Perceptron; Stress; Wire tool
• ISSN: 1877-7058; 1877-7058
• DOI: 10.1016/j.proeng.2017.10.659

The article presents a study of the piles stress state of a wire tool, which is a basic element in cladding systems. The technology consists from the deformation surface hardening processed simultaneously with thermomechanical coatings. The paper also presents a correct estimation of the wire tools piles durability and the ways to improve its fatigue endurance. It is considered how wire tool constructional parameters and cladding modes affecting the value of the operating maximum stresses in the piles while their sliding along the contact area with the workpiece. The calculation of the endurance limit is made by taking into account the maximum temperatures of flexible tool piles during sliding. The CAE-simulation results of wire tool piles stress state and the comparative analysis with the results of the calculation by the analytical method are presented. It is proposed to apply the methods of artificial intelligence in the field of pattern recognition to find the allowable region of the flexible instrument parameters and modes of the cladding. It was found that the use of a neural network and perceptron for modeling cladding operation provides the least learning error in parameters set classifying on areas of acceptable and unacceptable values for allowable stress criteria.