From particle size distribution to flow function coefficient in seconds. Physics-informed deep learning replaces weeks of shear cell testing.
Powder flowability is critical to pharmaceutical manufacturing, yet measuring it is slow, expensive, and impossible for blends that don't exist yet.
Shear cell measurements take 30-60 minutes per sample. Screening dozens of formulations can take weeks.
Each shear cell test consumes grams of expensive API or excipient that may be in limited supply during early development.
You cannot predict how a new blend will flow without physically making it. Computational screening of blend ratios is not possible with traditional methods.
Powmetrix combines a Bond Number physics model with a convolutional neural network trained on real shear cell data. Input a particle size distribution; receive a flow function coefficient (FFC) with confidence intervals.
Enter d10, d50, d90 percentiles or provide a full particle size distribution
Bond Number model provides a physics baseline; ResNet CNN learns residual corrections
Flow function coefficient with confidence interval and flowability risk classification
Instant FFC prediction from particle size percentiles or full PSD data with Monte Carlo uncertainty quantification.
Predict flowability across binary blend ratios. Visualize the FFC landscape before making a single blend.
Build formulations from your material library. Adjust weight fractions and predict blend flowability in real time.
Shared material database with role-based access. Your entire team works from the same validated data.
Measured vs. predicted FFC for 33 pharmaceutical excipients. Each point represents the average of triplicate shear cell measurements compared against the model prediction from PSD data alone.
Powmetrix is currently in early access.
Contact us for pricing
Create your organization account and start making predictions in minutes.
Create Your Account