US 11,837,333 B1
Simulation guided inverse design for material formulations
Bo Shuang, Missouri City, TX (US); Sukrit Mukhopadhyay, Midland, MI (US); and Fabio Aguirre Vargas, Lake Jackson, TX (US)
Assigned to Dow Global Technologies LLC, Midland, MI (US)
Filed by Dow Global Technologies LLC, Midland, MI (US)
Filed on Dec. 20, 2022, as Appl. No. 18/085,336.
Int. Cl. G16C 20/30 (2019.01); G16C 20/70 (2019.01); G06N 5/022 (2023.01)
CPC G16C 20/30 (2019.02) [G06N 5/022 (2013.01); G16C 20/70 (2019.02)] 10 Claims
OG exemplary drawing
 
1. A method of generating a hypothetical formulation library of chemical formulations, comprising generating hypothetical formulations by one or more of:
(a) directed combinatorics, comprising:
selecting one or more starter formulations from historical data;
performing substitution of one or more formulation components in at least one component class for each of the starter formulations; and
assigning concentration ratios to the formulation components within ranges established by the historical data; and
(b) constrained randomization, comprising:
preparing one or more template formulations comprising one or more component classes based on historical data;
perform randomized substitution of formulation components for each of the component classes of the template formulations;
randomly assign concentration ratios to the formulation components within ranges established by one or more constraints to produce one or more hypothetical formulations;
create one or more descriptors for the hypothetical formulations;
compare descriptor outputs from the hypothetical formulations with the range of descriptor outputs from the historical data; and
remove duplicate and outlier formulations; and
collecting the hypothetical formulations into the hypothetical formulation library.