Predicting Polymer Ductile-Brittle Transition via Impact Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods lack an efficient and cost-effective way to predict the ductile-to-brittle transition temperature (DBTT) of polymer compositions, requiring extensive experimentation and specialized equipment, which hinders the development of new polymer compositions for various applications.
Innovation Solution
A computer-implemented method using machine learning algorithms trained with impact curve data from training polymer compositions to predict the DBTT of test polymer compositions, allowing for rapid prediction within seconds to minutes based on a single experiment at a constant temperature, reducing the need for extensive experimental data and specialized equipment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If extensive experimentation is conducted to determine DBTT by measuring toughness over a wide range of temperatures, then measurement precision is improved, but loss of time and productivity deteriorate
Solution Approach 1:
Impact curves are measured in advance at a single reference temperature during material formulation. These preliminary measurements capture the full stress-strain behavior and energy absorption characteristics, which are then used by the machine learning model to predict DBTT without requiring subsequent temperature-sweep experiments.
Solution Approach 2:
The machine learning model creates a virtual copy of the temperature-dependent toughness behavior based on the single reference temperature impact curve. The model learns the relationship between impact curve features and DBTT from training data, then applies this learned relationship to predict DBTT for new compositions without physical temperature variation experiments.
2Measurement precision
If toughness measurements are performed over a wide range of temperatures, then measurement precision is improved, but device complexity and experimental effort worsen
Solution Approach 1:
The temperature variable is extracted from the experimental procedure. Instead of varying temperature during measurement, the method extracts temperature-dependent behavior information from the impact curve shape and energy absorption characteristics at a single reference temperature, using machine learning to infer DBTT without physical temperature cycling.
3Adaptability or versatility
If a machine learning model is trained with Raman spectral data, then prediction capability is improved for certain properties, but device complexity and cost worsen due to laser equipment requirements
Solution Approach 1:
The method uses inexpensive, readily available impact testing equipment instead of expensive specialized equipment like Raman spectrometers with lasers. The impact curves obtained from simple mechanical testing serve as sufficient input features for the machine learning model, eliminating the need for costly measurement apparatus while maintaining prediction capability.
Data Source
Figure 1
Figure 2
Figure 3a~3b
AI summary
The present invention relates to a computer implemented method for predicting the ductile-to-brittle transition temperature of a test polymer composition based on impact curves. Furthermore, a non-transitory computer readable storage medium is provided for tangibly storing computer program instructions capable of being executed by a processor, the computer program instructions defining the steps the aforementioned computer implemented method. Furthermore, the invention is directed to the use of impact curves and values indicative of the ductile-to-brittle transition temperatures of polymer compositions for training machine learning algorithms.