The Digital Transformation of Product Formulation: Concepts, Challenges, and Applications for Accelerated Innovation

In competitive manufacturing industries, organizations embrace product development as a continuous investment strategy since both market share and profit margin stand to benefit. Formulating new or improved products has traditionally involved lengthy and expensive experimentation in laboratory or pilot plant settings. However, recent advancements in areas from data acquisition to analytics are synergizing to transform workflows and increase the pace of research and innovation. The Digital Transformation of Product Formulation offers practical guidance on how to implement data-driven, accelerated product development through concepts, challenges, and applications. In this book, you will read a variety of industrial, academic, and consulting perspectives on how to go about transforming your materials product design from a twentieth-century art to a twenty-first-century science. Presents a futuristic vision for digitally enabled product development, the role of data and predictive modeling, and how to avoid project pitfalls to maximize probability of success Discusses data-driven materials design issues and solutions applicable to a variety of industries, including chemicals, polymers, pharmaceuticals, oil and gas, and food and beverages Addresses common characteristics of experimental datasets, challenges in using this data for predictive modeling, and effective strategies for enhancing a dataset with advanced formulation information and ingredient characterization Covers a wide variety of approaches to developing predictive models on formulation data, including multivariate analysis and machine learning methods Discusses formulation optimization and inverse design as natural extensions to predictive modeling for materials discovery and manufacturing design space definition Features case studies and special topics, including AI-guided retrosynthesis, real-time statistical process monitoring, developing multivariate specifications regions for raw material quality properties, and enabling a digital-savvy and analytics-literate workforce This book provides students and professionals from engineering and science disciplines with practical know-how in data-driven product development in the context of chemical products across the entire modeling lifecycle.

Table of Contents:
Section 1: Getting Started. 1. The Digital Transformation of R&D Labs. 2. Product Formulation Fundamentals. 3. Defining a Successful Predictive Formulation Project. Section 2: Preparing Your Data. 4. Challenges with Formulation Datasets. 5. Feature Engineering: Enhancing Your Data with Descriptors. 6. Machine Learning for Analysis of Structural Characterization. Section 3: Predictive Modeling. 7. Machine Learning Techniques for Predicting Properties of Formulations. 8. Modeling of Product Formulations Using a Latent Variable Approach. 9. Gaining Trust in Your Model. Section 4: Optimization and Inverse Design. 10. Introduction to Formulation Optimization. 11. Adaptive Experimental Design. 12. Inverse Design via PLS Model Inversion. Section 5: Case Studies and Special Topics. 13. Case Studies. 14. Special Topics. 15. Conclusion.

About the Author :
Alix Schmidt is a senior data scientist in Dow’s Core R&D Information Research team in Midland, Michigan. Alix earned a BS in chemical engineering at the University of Illinois Urbana–Champaign in 2009 and then joined Dow Corning initially as a process research engineer. Since then, Alix has held a variety of roles at Dow Corning and Dow and completed an MS in data science at Northwestern University. Alix has experience with polymer process research, high-throughput research, machine learning for manufacturing troubleshooting, and data-driven product development. Her interest and experience in materials informatics allow her to lead technical data science strategy at Dow, and she has presented and chaired at the AIChE spring meeting on this topic. Kristin Wallace earned a BS in chemical engineering (2006) and an MS in applied science (optimization focus) (2008) at McMaster University. She has worked on a variety of analytics projects since joining ProSensus Inc. in 2018 as a project engineer in Burlington, Ontario. Her particular interest in product formulation using projection to latent structures (PLS) has led her to be involved with related consulting projects, contributing to the development of FormuSense (commercial software), authoring blogs and magazine articles, as well as presenting and chairing at several AIChE spring meetings. Prior to working at ProSensus, she spent five years designing and troubleshooting non-ferrous electric arc furnaces.