Related Experiment Video
Updated: Oct 24, 2025
08:43
Fused Filament Fabrication FFF of Metal-Ceramic Components
Published on: January 11, 2019
17.5K
Using a material database and data fusion method to accelerate the process model development of high shear wet
Zheng Wang1, Junjie Cao1, Wanting Li1
1Department of Chinese Medicine Informatics, School of Chinese Materia Medica, Beijing University of Chinese Medicine, No.11, North Third Ring East Road, Beijing, 100029, People's Republic of China.
Scientific Reports
|August 14, 2021
Summary
Data fusion enhances high shear wet granulation (HSWG) process modeling. Combining literature and lab data significantly reduces prediction errors for granule size, saving time and costs in oral solid dosage manufacturing.
Area of Science:
- Pharmaceutical Technology
- Chemical Engineering
- Process Modeling
Background:
- High shear wet granulation (HSWG) is a key process in oral solid dosage (OSD) manufacturing.
- Understanding and controlling the complex HSWG process requires robust process modeling.
- Existing models often struggle with diverse formulation data and varying equipment scales.
Purpose of the Study:
- To develop a data fusion methodology for creating a comprehensive formulation-process-quality model for HSWG.
- To improve the accuracy of predicting granule size by integrating diverse datasets.
- To demonstrate the benefits of data fusion in accelerating HSWG formulation development and reducing experimental costs.
Main Methods:
- Data fusion of experimental HSWG data from literature and laboratory sources.
- Utilizing a material database and material matching to address incomplete physical characterization.
- Employing dimensionless parameters to reconstruct process variables across different granulator scales.
- Applying Principal Component Analysis (PCA) for exploratory analysis of input material properties.
- Developing Partial Least Squares (PLS) regression models for median granule size prediction.
Main Results:
- Exploratory PCA revealed significant diversity within the formulation data library from various literature sources.
- PLS models incorporating both literature and laboratory data showed significantly reduced prediction errors for median granule size compared to literature-only models.
- The benefits of data fusion were most pronounced at low liquid-to-solid ratios.
- The methodology successfully compensated for incomplete material characterization and scaled process variables.
Conclusions:
- The proposed data fusion methodology is effective for building robust HSWG models.
- Integrating a small amount of laboratory data can substantially improve the predictive power of multivariate models.
- This approach offers significant advantages in saving experimental time and cost for HSWG formulation and process development.
