Use of artificial neural networks for determining optimum thread forming speeds for thread forming fasteners

摘要

Purpose - Develops a neural network based system for optimum assembly speeds using thread-forming fasteners. Design/methodology/approach - Uses a three layer neural network to optimise thread forming speeds based on thread diameter and pitch and the total number of thread coils. Findings - The method demonstrates savings in energy and reduction in torque values of 20-30 per cent. Research limitations/implications - Provides a method that works even when less experimental data are available. Practical implications - The method should provide a higher quality and reliability and allow thread-forming fasteners to be used in new application areas. Originality/value - Provides an efficient and less labour intensive method for insertion speed optimisation.