Method and apparatus for the rapid discovery and design of polymerizations
AssigneeNew York University
InventorsRyan Lee Hartman, Benjamin A. Rizkin
Life of this patent
Estimating the term…
Abstract
A reactor system includes at least one reactant provided to perform a reaction. The system includes one or more sensors configured to detect sensor data regarding the reaction. The system includes processing circuitry configured to receive the sensor data from the one or more sensors, apply one or more machine learning models to the sensor data to generate a measurement regarding at least one of the reaction or an activity of at least one catalyst used to perform the reaction, and control at least one of a temperature of the reactor, a flow rate of the at least one reactant, or a concentration of the at least one reactant responsive to the measurement.
Claims20
A system, comprising:
- a flow reactor, comprising:
- at least one reactant provided to perform a reaction;
- one or more sensors configured to detect sensor data regarding the reaction;
- and
- processing circuitry configured to:
- receive the sensor data from the one or more sensors;
- apply one or more machine learning models to the sensor data to generate a measurement regarding at least one of kinetics of the reaction or an activity of at least one catalyst used to perform the reaction;
- and
- control at least one of a temperature of the flow reactor, a flow rate of the at least one reactant, or a concentration of the at least one reactant responsive to the measurement.
A method comprising:
- receiving, by processing circuitry, sensor data regarding a reaction from one or more sensors that detect the sensor data by monitoring a flow reactor;
- applying, by the processing circuitry, one or more machine learning models to the sensor data to generate a measurement regarding at least one of kinetics of the reaction or an activity of at least one catalyst provided in the flow reactor;
- and
- controlling, by the processing circuitry, at least one of a temperature of the flow reactor, a flow rate of at least one reactant of the flow reactor, or a concentration of the at least one reactant responsive to the measurement.
A method, comprising:
- providing training data as input to a machine learning model, the training data comprising (1) a measure of catalytic activity of a reaction between at least one catalyst and at least one reactant and (2) at least one output parameter regarding at least one product generated by the reaction;
- causing the machine learning model to generate candidate output responsive to the training data, the candidate output comprising at least one candidate output parameter;
- evaluating the candidate output using an objective function;
- modifying the machine learning model responsive to the evaluation.
Dependencies are inferred from the claim text.