Pulp Mill Reduces Brightness Variation by 80% with Mt. Fuji
Haber deployed Mt. Fuji®, its industrial AI platform for predictive process optimization, at a pulp mill to stabilize pulp brightness and optimize bleach chemical dosing. By predicting brightness in real time and prescribing peroxide and caustic dosing, Mt. Fuji significantly reduced brightness variability, lowered chemical consumption, and improved pulp quality without disrupting mill operations.
The Challenge
Maintaining consistent pulp brightness is one of the biggest challenges in BCTMP production. Variations in wood chip quality, bleach chemical performance, process conditions, and brightness sensor accuracy often force operators to overdose bleaching chemicals as a safety margin.
The mill's bleaching process consisted of pre-bleaching, refining, bleaching, and washing stages, each influencing final pulp brightness. Manual process adjustments based on delayed laboratory measurements resulted in:
- Brightness deviations of ±5 ISO points
- High bleach chemical consumption
- Elevated peroxide residuals
- Variable pulp strength
- Limited visibility into process performance
Haber's Approach
Real-time brightness prediction with sensor inputs and lag-time adjustments.
The platform recommends the optimal peroxide and caustic dosage based on live alkalinity ratios and process conditions, helping maintain target brightness while minimizing chemical consumption.
Dosing recommendations dynamically adapt to incoming Kappa number, pulp consistency, temperature, and other process variations, ensuring stable bleaching performance despite changing operating conditions.
Self-learning AI models continuously adapt to changing wood chip quality, seasonal variability, and process drift, improving prediction accuracy and optimization performance over time.
Evidence-Backed Results
| Metric | Before | After | Change |
|---|---|---|---|
| Final tower brightness deviation | ±5 ISO points | ±1 ISO point | 80% reduction |
| Bleach chemical costs | Baseline | 10% Lower | 10% reduction |
| Bleach chemical residuals | High | Significantly reduced | Improved chemical utilization |
Business Impact
By replacing reactive bleaching control with predictive AI-driven optimization, Mt. Fuji enabled operators to maintain target brightness using only the required amount of peroxide and caustic.
Conclusion
Mt. Fuji transformed brightness control from a reactive laboratory-based workflow into a predictive AI-driven optimization system. By forecasting pulp brightness and prescribing peroxide and caustic dosing in real time, the platform stabilized product quality, reduced chemical costs, and enabled more efficient BCTMP production.
