Probabilistic Design for Optimization and Robustness for Engineers

12 Verification and validation

12.1 Introduction

Throughout this book, we have discussed techniques to develop optimal, robust products. In some cases, the primary focus has been on optimization techniques relying on engineering models, although we have also discussed using designed experiments and logistic regression analysis. Regardless of the approach, any design optimization effort requires confirmation.

As done in software engineering (IEEE, 1998), we may make a distinction between verification and validation. Verification involves the review and evaluation of designs, documents, and plans to ensure we have fulfilled the specified design requirements. In other words, “Does the design satisfy the specifications?” Usually, this is done prior to physical evaluation using actual production intent processes. It may be performed using design checklists, walkthroughs, meetings, and simulations. Typically, it is a virtual or paper exercise.

Validation, on the other hand, involves the review and physical evaluation of the actual product. It involves physical testing of the product relative to customer and functional requirements and requires acceptance or approval by the next process customer or end user. It answers the question, “Does the design satisfy the customer?” In some cases, final validation may not be completed until after a product has been produced using the actual production process and is in use by customers.

As per the above operational definitions, verification ...