Innovation Requires a Balance Between Sustainability and Performance
The future of sustainable innovation in the coatings industry
The coatings industry faces a growing challenge: how to innovate while meeting susta inability goals and without compromising on performance. In this interview, Erik Sapper from California Polytechnic State University shares his insights on balancing these priorities, the role of emerging technologies, and what the future holds for research and development in coatings science.
Question 1
Question 2
Question 3
Question 4
What realistic role does AI and simulation currently play in coatings development?
Modeling and simulation have a decades-long track record of helping to characterize and understand various aspects of coatings development, including resin synthesis, polymer property prediction, and the discovery of formulation rules and best practices. Simulation has also been used extensively to understand and improve production environments from a chemical engineering, manufacturing, or delivery perspective.
Artificial intelligence and machine learning broaden the impact of using a data-driven or algorithmic approach to the coatings development process. I think it’s important to appreciate the pattern-finding ability of these algorithms and tools. Given enough quality data, these computational tools can help find patterns, discover heuristic rules of thumb or best practices, and can optimize highly dimensional chemical and manufacturing spaces, much more quickly than a team of humans alone can.
Where do predictive models add the most value, and where is human expertise still essential?
Extracting value from a predictive model can be an uphill endeavor. If the model only lives on your laptop or in your slide deck, it’s probably not imparting real organizational change. A decent model that is well-deployed across the company is infinitely better than a great model that doesn’t get used by anyone. So, deployment is key. With this comes the responsibility of training users, having good data infrastructure in place, and creating a culture of embracing data, whether that data is good, confirmatory data, or if it is results from a failed experiment; all of this data needs to be captured to drive these gains.
In practice, I still see many predictive models being used to design new resins and new additives. I’m seeing more on applying predictive models to formulation design, discovery, and optimization, but much of that work still relies on iterative improvements over past products or starting point formulations.
Which areas of coating production are benefiting most from automation today?
Synthesis, testing and characterization, and, of course, production and application environments, all benefit from automation. The coming decades will see real increases in automating what used to be considered the more ad hoc processes in coatings development, experiments like one-off resin syntheses, ladder studies, and standardized tests. The stuff that used to seem like busy work, or good work for summer interns or entry level employees, will slowly begin to be automated throughout the industry. That doesn’t mean those people aren’t needed; instead, we’ll get to use their expertise in solving more exciting and challenging chemical problems.
What prerequisites must be met for successful automation in coating manufacturing?
To successfully begin automating your coating manufacturing process, you have to understand the problem you’re trying to solve and the exact process you are trying to automate, especially from a business perspective. If you automate something that doesn’t benefit from automation, that’s not value added.
You need to know how data flows through your lab. How it’s generated, how it is used, and what decisions the data informs or enables. When you’re automating a lab you’re automating physical processes, sure, but you’re also automating decisions. Getting clear on those decisions will help the automation process be as frictionless as possible.
