Thouheed Abdul Gaffoor
CEO & Co-Founder, Basetwo AI
Thouheed Abdul Gaffoor is the Co-Founder and CEO of Basetwo AI, a company that deploys digital twins to optimize scale up and manufacturing in the chemicals industry. Thouheed previously served as the Head of AI at Autodesk, and has spent over a decade leading global deployments of AI solutions across multiple industries, working with Fortune 500 manufacturers to improve process efficiency, quality, and sustainability. With a background in engineering from the University of Waterloo, he specializes in applying physics-informed machine learning to solve complex industrial challenges.
Digital Twins for Optimized Quality Control in Hot Melt Adhesive Manufacturing
Hot melt adhesives are a critical segment of the specialty chemicals industry, serving high-performance applications in packaging, hygiene, automotive, electronics, and construction. These materials are formulated from blends of base polymers, tackifiers, waxes, plasticizers, and additives that must be fully melted and homogeneously dispersed to achieve consistent rheological and mechanical properties. Among all quality attributes, melt viscosity is the most critical, directly impacting performance, stability, and downstream processability. Even small viscosity deviations can cause coating defects, poor adhesion, excess material usage, or unplanned downtime.
Achieving target viscosity in batch manufacturing is a complex thermo-mechanical challenge. During charging and heat-up, solid polymers and resins undergo temperature-dependent phase transitions while viscosity evolves nonlinearly with temperature and composition. Heat transfer is constrained by jacket design, agitation, and the low thermal conductivity of partially melted masses. As melting progresses, changes in torque, shear, and convection further couple thermal and rheological behavior, creating internal temperature gradients and transient melt fronts that cannot be directly observed.
Operators often rely on conservative setpoints and long cycle times to ensure full melting, leading to excess energy use, higher raw material costs and extended cycle times.
This presentation introduces a hybrid digital twin framework for hot melt mixing that integrates first-principles heat transfer, melting dynamics, and rheological models. The twin estimates melt state, viscosity evolution, and mixing progression using live data, enabling proactive control of temperature to reliably achieve target viscosity while improving efficiency and cycle time.
Co-authors:
Emma McGlade
Breakout Session XIII – Data-Driven Innovation in Adhesives – 18 September 2026 – 11:30 – 12:00 – Room Fleming – F3