Systematic method for axiomatic robustness-testing (SMART)

Abstract

SMART (Systematic Method for Axiomatic Robustness-Testing) is a method for the development of robust and reliable products. It combines elements from the robust design methodology with a holistic approach by using Axiomatic Design (AD) and the Taguchi Method (TM). These two methods were established and expanded by N.P. Suh [1990] (AD) and G. Taguchi [1949] (TM).

SMART is based on the chronological sequence of the four phases of the Product Development Process (planning, conception, design and development) according to the VDI Guideline 2221. Using this chronological basis, the three process steps (System, Parameter and Tolerance Design) of the Taguchi Method are classified and integrated accordingly. The AD method is applied to the systematic examination of the robustness of designs.

During the conceptual stage, one or more designs are generated by means of AD. AD also helps analyze the design’s complexity from the perspective of possible design modifications, thus assuring robust solutions.

If a design has already been generated but needs improvement as things developed, AD is used as well. The design may not necessarily be changed in its basic structure but is examined in terms of its complexity. The results of AD support the setup of the P-Diagram according to Taguchi either after the conceptual stage or the design stage of the product.

The following step is the Design of Experiments (DoE) of the product’s design parameters and noise factors that occur during its utilization. Testing may either be carried out by virtual or real tests. After analyzing the results of the tests, the design should be optimized accordingly in order to increase the robustness. A predicted reliability determination is possible as well.

The last step is the adjustment of the tolerances of the design for cost optimization purposes. After a final robust design has been established, the actual durability and reliability of the design can be determined on the basis of reliability testing using Design for Reliability (DFR) methods.

Basically, SMART can be used both in the initial stages as well as in the more developed stages of the development process.