Adaptive Quality Estimation From Density and Temperature Data

The patent replaces expensive NIR spectroscopy equipment with a cost-effective solution using simple density and temperature measurements combined with regression models. This approach uses inexpensive, readily available measurement tools rather than costly specialized equipment, achieving acceptable measurement precision at significantly lower investment and operational costs.

The patent substitutes complex optical measurement systems (NIR spectroscopy) with simpler physical measurement approaches based on density and temperature measurements. By using regression models to infer quality parameters from these basic measurements, the system replaces expensive mechanical/optical equipment with affordable sensors and computational methods.

The patent transforms static fixed-parameter regression models into dynamic adaptive models that continuously update their parameters using new measurement data. This allows the model to adapt to changing feedstock properties and process conditions, maintaining measurement precision over time without requiring complete re-modeling, thus balancing low costs with sustained accuracy.

The patent implements a feedback mechanism where new density and temperature measurements are continuously used to update the regression model parameters. This closed-loop approach ensures the model remains accurate by incorporating latest data, preventing model obsolescence while maintaining cost-effectiveness compared to expensive alternative methods.