Innovation Award

About the award

This Award is for the company or operating unit which can provide the most successful examples of innovation – turning ideas into money – in an area of its business that helps solve today’s societal challenges.

Dow Polyurethanes predictive intelligence entry demonstrated real quantified benefits to the company, vastly cutting development time and saving resources. Judges felt this entry also had fantastic transferability potential to encourage greater uptake of AI tools in the sector.

Dow’s Predictive Intelligence (PI) platform is revolutionising the polyurethane (PU) formulation development process by leveraging AI, advanced data analytics, and machine learning. By organising and enriching over 20,000 historical and commercial formulations, Dow has created a system that accelerates the time to market for innovative and sustainable solutions.

 PI’s patented machine learning models predict the properties of PU formulations and suggest novel compositions tailored to specific customer needs, reducing formulation discovery time from months to weeks. This data science ecosystem, using structured query language databases, enhances customer experience and drives business growth.

 Sustainability is a key focus of the PI platform, as demonstrated by its application in the RENUVA™ Mattress programme. Researchers can search by properties, material hazards, and commercial availability, facilitating the development of sustainable materials. While real-world validation is still necessary, PI enables more accurate and efficient formulation selection.

 Deployed globally across Dow’s polyurethane business, PI delivers multimillion-dollar customer project wins by delivering better, safer, and recyclable materials. Dow has globalised the PI capability by designing a user-friendly, cloud-based interface for PU formulators worldwide. This tool accelerates multivariate optimisation and identifies ideal candidates for customer needs across various polyurethane business segments.

 Historically, formulators in the polyurethane foam business relied on experience and intuition, necessitating manual searches and multiple iterations of experiments to optimise foam formulations. This process could take months to years. Now, PI enables real-time searches of the formulation space, significantly reducing the number of experiments needed, and producing solutions in weeks.

 The tool empowers industry experts with cutting-edge formulation search and patented AI-driven modelling capabilities, facilitating the innovation of PU foam formulations. Whether identifying PU foams with specific performance criteria or substituting hazardous components with sustainable materials, PI leads to quicker, safer, and renewable solutions in the polyurethane foam formulation business.

 Dow can combine this technology with its deep material expertise to innovate sustainable and cost-effective solutions in the PU space. Scientists have already identified customer-ready formulations in the first search and verification, even outside North America. Leveraging a dataset of over 20,000 foam formulations and proprietary AI models to select experiments that align perfectly with application requirements represents a revolutionary advancement in the chemical industry. It enhances the speed and efficiency of innovation, particularly for recycled polyurethane materials.

 The PI solution architecture leverages the latest cloud technologies. Dow’s data engineers and architects developed a database and data pipeline using Azure Data Factory and Azure Synapse, allowing easy dataset updates and secure data queries from the user interface. Dow’s managed Azure Kubernetes Cluster deploys the Predictive Service as a scalable microservice. The development team created an Angular application and .NET Web API to communicate with the Predictive Service, presenting users with a modern web app that enables formulation searches, real-time predictive experiments, and results export to Excel for direct experimentation.