The Design of Experiments as a Methodological Framework for the Improvement of Manufacturing Processes

Introduction

The manufacturing industry faces constant challenges, mainly in the manufacture of innovative and high-quality products, due to the current conditions of the global market, and the high competitiveness of the organizations. Based on the concept of manufacturing, which is defined as: “the transformation of materials and information into goods for the human needs satisfaction” (Chryssolouris, 2006), the processes for product development are increasingly complex since the products become more versatile to meet the needs of massive customization (Ong, Yuan, & Nee, 2008). Therefore, it is necessary to make important changes to the manufacturing processes by designing or re-designing them, in order to improve the quality of products, reduce costs and cycle times, change materials, modify methods, design innovative products and highly competitive in the market, among others (Gutiérres Pulido & De la Vara Salazar, 2012).

Different manufacturing phases require experimentation in order to be able to face critical factors such as optimization of available resources, of the execution time and the process performance, as well as variability control, to assure the quality of the product and/or service produced (Tanco, Viles, Ilzarbe, & Álvarez, 2007). A powerful methodological framework that allows to create a multidimensional design space and its interactions between factors that might affect the performance and quality of the process is known as Design of Experiments (DoE) (Davim, 2012). This framework provides us with objective evidence that responds to diverse questions in the different stages of process development, clarifying uncertain aspects that lead us either to the improvements implementation or problem solutions with different complexity levels (Gutiérres Pulido & De la Vara Salazar, 2012). DoE can be understood as a practical methodology, a tool that is composed of a set of techniques with well-defined application phases, which are described below (Murray et al., 2016; Weissman & Anderson, 2015):

• 1.

  Planning, this phase begins with the recognition of a need and/or problem, as well as the definition of the objectives to be achieved, requires a good level of process knowledge, or an evaluation and comprehensive analysis that allows to identify the factors (variables) involved in the process, as well as, define the expected responses and list all possible interactions between them. This exercise contributes to the construction of an environment where the culture of decision-making is fostered based on experimental design techniques, rather than those based on trial and error, providing greater certainty to the decision making process.

• 2.

  Design, in this phase the most appropriate DoE is chosen, based mainly on the number of factors that must be studied, the different levels of detail, as well as the objectives to be achieved.

• 3.

  Conducting, refers to the experimentation, execution and the evaluation of the results obtained.

• 4.

  Analysis, in this stage the results are analyzed and interpreted, in order to obtain significant conclusions for the decision making process, to reach objectives such as parametric design or to identify the process variables that critically affect its performance, the influence or sensitivity due to variability of the external components, the identification of the different parametric levels that produce an optimal performance, and finally, to identify if the improvement or implementation can be executed in a real environment.