Chemical Reactor Control Settings Using Generative ML for Polymer Discovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for synthesizing polymers are laborious and rely heavily on human discretion, lacking efficiency in exploring optimal structure-activity relationships, as they involve manual prediction of chemical structures, synthesis schemes, and reactor control settings.
Innovation Solution
The use of generative machine learning models, such as variational autoencoders (VAE) and gain adversarial networks (GAN), to autonomously generate recommended chemical reactor control settings for the synthesis of polymers, based on past operation data, enabling efficient and autonomous control of chemical reactors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional manual methods are used for polymer synthesis, then human discretion can be applied to predict chemical structures and synthesis schemes, but the process is laborious and time-consuming
Solution Approach 1:
The patent replaces manual mechanical prediction processes with an automated machine learning system. The ML model processes historical chemical reactor data to generate predictions for chemical structures, synthesis schemes, and reactor control settings, eliminating the need for manual human analysis and significantly reducing discovery time.
Solution Approach 2:
The system enables autonomous self-service by allowing the machine learning model to independently generate and optimize polymer synthesis parameters without continuous human intervention. The model learns from historical data and automatically produces recommendations for new polymer discoveries, reducing labor requirements.
2Reliability
If conventional manual methods are used for synthesizing polymers, then human expertise can guide the process, but the efficiency in exploring optimal structure-activity relationships is low
Solution Approach 1:
The system performs preliminary actions by pre-training the machine learning model on extensive historical chemical reactor data before actual polymer discovery. This preliminary training enables the model to quickly generate reliable predictions during operation, combining the benefits of pre-computed knowledge with rapid exploration of new polymer possibilities.
Solution Approach 2:
The patent introduces a computational dimension to polymer discovery by using machine learning algorithms that can process and analyze multidimensional chemical data simultaneously. This enables efficient exploration of structure-activity relationships across multiple parameters at once, something that manual methods struggle to achieve.
3Loss of time
If generative machine learning models are used to autonomously generate control settings, then time and labor are reduced, but the system requires processing vast amounts of chemical reactor data
Solution Approach 1:
The patent segments the complex data processing task into distinct functional modules: data collection from historical reactors, data preprocessing and cleaning, feature extraction, model training, and prediction generation. This segmentation makes the complex system more manageable and allows each component to be optimized independently.
Solution Approach 2:
The system introduces intermediate data structures and processing layers between the raw historical data and the final predictions. These intermediaries include processed training datasets, extracted features, and model representations that bridge the gap between vast historical data and actionable control settings, reducing the computational burden.
Data Source
AI summary
Techniques regarding autonomously controlling one or more chemical reactors using generative machine learning models are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a model component that can build a generative machine learning model based on training data regarding a past chemical reactor operation. The generative machine learning model can generate a recommended chemical reactor control setting for experimental discovery of a polymer.