New Algorithm (RABAL) to address confounding problems in Taguchi parameter Design--A Practical Study

Abstract

It is essential to operate all manufacturing processes at optimum levels to stay competitive in the global environment. Optimum prediction in manufacturing process involves lot of experimentation as it involves time and cost. Taguchi Parameter Design (PDE) is a design where optimum prediction is achieved in lesser experimental runs subject to the condition that there is no confounding effect. Confounding effect is a drawback of Taguchi PDE due to which many Taguchi experiments fail in confirmation tests at desired confidence level. RABAL Algorithm is 12 step structured approach, which uses Orthogonality concept and linear graph for Factors and Interaction allocation and uses certain rules based on statistical concepts by which the accurate result of Taguhi parameter design experiment is ensured. Sleeve synchronizer 1&2 hard turning is a bottleneck operation in a gear manufacturing line taken for research optimized for better cycle time using Taguchi PDE and RABAL algorithm. Feed rate in first and second cut is observed as significant factors at 95% confidence level and Optimal parametric setting is identified. Accuracy of results is confirmed through confirmation tests. RABAL Algorithm can be used along with Taguchi PDE to minimize the confounding effects there by accuracy of experimental results are improved and results are achieved right first time.