Multi-Phase Material Property Prediction With Phase-Aware AI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional AI methods for predicting material properties assume a single phase, leading to inaccurate predictions when materials undergo phase changes due to environmental or usage variations, particularly in lithium-ion secondary batteries where the cathode material transitions to multiple phases during charging and discharging.
Innovation Solution
A system and method using AI models to input material and phase information to generate feature data, which is then processed through multiple AI models to accurately predict properties of materials with multiple phases, utilizing pre-learned feature data as a representative vector without needing separate phase information for each prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional single-phase AI prediction method is used, then the prediction process is simple, but the prediction accuracy deteriorates when materials undergo phase changes
Solution Approach 1:
The patent segments the prediction process into multiple independent AI models, each dedicated to predicting properties of a specific material phase. Instead of using a single general model, the system divides the prediction task across multiple specialized models corresponding to different phases (e.g., cubic phase, spinel phase, rock-salt phase), thereby improving accuracy for each phase while maintaining manageable complexity through modular architecture
Solution Approach 2:
The patent changes the parameter representation by introducing phase identification as an additional input parameter to the AI system. The model accepts both material composition parameters and phase parameters, allowing it to adapt its prediction behavior based on the identified phase. This parameter expansion enables the system to handle phase transitions while maintaining prediction accuracy
2Measurement precision
If multiple AI models are used to account for different phases, then prediction accuracy improves, but system complexity increases
Solution Approach 1:
The patent creates a universal AI prediction system that can handle multiple phases through a standardized architecture. Each phase-specific model follows the same input-output structure and can be trained independently, allowing the system to universally apply the same methodology across different material phases without requiring complex customizations for each phase
Solution Approach 2:
The patent introduces a phase identification module as an intermediary that determines which AI model should be used for prediction. This mediator analyzes material properties and phase characteristics, then routes the prediction task to the appropriate phase-specific model, thereby managing system complexity through intelligent task distribution rather than requiring all models to operate simultaneously
3Reliability
If traditional material testing methods are used, then comprehensive verification is achieved, but development time and cost increase significantly
Solution Approach 1:
The patent performs preliminary prediction of material properties using AI models before actual material synthesis and testing. By predicting properties of different phases and their mixtures in advance, the system identifies optimal material compositions and phase ratios, allowing researchers to focus experimental verification only on the most promising candidates rather than testing all possible combinations
Solution Approach 2:
The patent creates virtual copies of material systems through AI simulation. Instead of physically synthesizing and testing numerous material variants, the system generates virtual material models with predicted properties, allowing comprehensive verification in silico before moving to physical experimentation, thereby reducing time and resource requirements
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A system, method, and program predict the properties of a material having a multi-phase. The system for implementing an AI model of predicting the properties of a material having a multi-phase includes memory configured to store instructions that are executable; and one or more processors configured to execute the instructions to perform operations comprising: inputting first material information, which includes information regarding the material, into a first AI model to output first feature data; inputting first phase information, which includes information regarding a first phase of the material, into a second AI model to output first phase feature data; and inputting second phase information, which includes information regarding a second phase of the material, into the second AI model to output second phase feature data; and wherein the first feature data includes information regarding the properties of the material according to the multi-phase of the material.