Optimization of actuator performance using robust engineering and feature selection methodologies : A case study

Purpose – The purpose of this study is to demonstrate the variation between the set torque and the actual torque at which the actuator trips can be minimized using Taguchi's robust engineering methodology. The paper also aims to demonstrate the application of feature selection approach for the identification of insignificant effects in unreplicated fractional factorial experiments. Design/methodology/approach – The methodology used was design of experiments with the set torque as the signal factor and the tripping torque as response variable. The compounded noise factor was identified based on the type of operations and load variation, which are not under the manufacturer's control. The effect of five control factors (with two levels each) and two interactions were studied. The experiments were designed using L8 orthogonal array. Findings – The result showed that the factors spring height, spring thickness, star washer position and the interaction between drive shaft length and spring height play a significant role in actuator performance. The implementation of the optimum combination of factors resulted in improving the overall capability indices, Cp from 0.52 to 2.12 and Cpk from 0.4 to 1.67. Practical implications – This study provides valuable information to actuator manufacturers on optimizing actuator performance. Originality/value – To the best of the author's knowledge, no study has been conducted using Taguchi's robust engineering methodology to optimize actuator performance. In addition, no attempt has been made in the past to identify the insignificant factors and interactions using feature selection approach for unreplicated fractional factorial experiments.