Specialty Paper Mill Reduces Wet Tensile Variability by 73% and Cuts WSR Consumption by 15% Using Mt. Fuji AI
Specialty Paper Mill Reduces Wet Tensile Variability by 73% and Cuts WSR Consumption by 15% Using Mt. Fuji AI
Haber deployed its Mt. Fuji® industrial AI platform at a specialty paper mill to stabilize wet tensile strength and optimize Wet Strength Resin (WSR) consumption. By combining real-time process signals with laboratory quality measurements, Mt. Fuji predicted wet tensile performance and continuously prescribed optimal WSR dosage. The closed-loop AI system significantly reduced product variability while lowering chemical consumption without compromising product quality.
The Challenge
Wet tensile strength is one of the most critical quality parameters in specialty paper production. Under conventional control, mills rely on reactive adjustments after laboratory measurements become available. Variability in furnish characteristics and wet-end operating conditions forces operators to overdose Wet Strength Resin (WSR) as a safety margin, increasing chemical costs while still producing inconsistent product quality.
At this mill, wet tensile variation reached a standard deviation of 0.46 N/15 mm, requiring an average 27.28 kg/t of WSR to maintain specification.
Haber's Approach
Mt. Fuji continuously collected wet-end process data including flow, stock consistency, retention, pH, furnish properties, and other production variables alongside laboratory wet tensile measurements.
Machine learning models established relationships between process conditions and final wet tensile strength, accurately predicting product quality before laboratory results became available.
The platform generated optimal Wet Strength Resin dosage recommendations that maintained target wet tensile strength while minimizing chemical consumption.
Models continuously adapted to changing furnish composition, seasonal variation, and process drift, ensuring sustained optimization over time.
Evidence-Backed Results
| Metric | Before | After | Change |
|---|---|---|---|
| Wet tensile variability (SD) | 0.46 N/15 mm | 0.12 N/15 mm | 73% reduction |
| Average wet tensile | 10.28 N/15 mm | 10.49 N/15 mm | Maintained target quality |
| WSR consumption | 27.28 kg/t | 23.42 kg/t | 15% Reduction |
Business Impact
Mt. Fuji enabled operators to shift from reactive chemical dosing to predictive process control. Instead of compensating for uncertainty through excess resin usage, the AI platform anticipated quality deviations and prescribed the minimum WSR dosage required to consistently achieve target wet tensile strength.
For a reference 350,000 TPY specialty paper mill, the improvement translates into approximately $4.4 million in annual economic impact, driven primarily by chemical savings, reduced quality losses, faster grade stabilization, and fewer customer quality credits.
Conclusion
By combining predictive quality models with AI-driven dosage optimization, Haber helped the mill reduce wet tensile variability by 73% while lowering Wet Strength Resin consumption by 15%. The deployment demonstrates how closed-loop industrial AI can simultaneously improve product consistency and reduce operating costs in specialty paper manufacturing.
