Title: Robust screening of cake product characteristics by the Taguchi method

Authors: Marina J. Pouliou; George J. Besseris

Addresses: TEI Piraeus, Methonis 8-10, Athens, Greece ' School of Engineering and Science, University of the West of Scotland, High Street, Paisley PA1 2BE, Scotland

Abstract: The aim of this work is to present an application of Taguchi methods for robust design in an industrial setting in order to identify the critical factors affecting the process of baking a popular cake product. By using proven quality methods, such as design of experiments, finding the most significant factors that affect the baking process of a cake product is organised by experimenting with trials offline. This is to be succeeded by minimising variability in the food engineering process while reducing the percentage of defective cake produced in the production line. For this reason, a three-level, four-factor orthogonal array according to the Taguchi method was utilised to plan the experiments. The two investigated responses, the cake weight and surface peak, are to be analysed in separate counts. The design parameters which were taken into account in this experimental design are the baking duration, the baking temperature and two relative sequential mixing speeds regarding the cake product blending phase. The analysis of the results using the analysis of variance treatment allowed rapid screening of the experimental design data leading to the appropriate process controls for an optimal cake product.

Keywords: DOE; design of experiments; ANOVA; analysis of variance; food engineering; quality improvement; orthogonal array; SNR; signal-to-noise ratio; Taguchi method; robust design; baking process; cakes; process control; baking duration; baking temperature; mixing speeds; cake product blending; rapid screening.

DOI: 10.1504/IJISE.2013.053737

International Journal of Industrial and Systems Engineering, 2013 Vol.14 No.2, pp.207 - 229

Published online: 27 Dec 2013 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article