Transferable Hybrid Prognostics for Degradation Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for developing degradation models in engineering systems are hindered by complexity, nonlinearity, and the need for abundant historical data, limiting their applicability across different domains due to domain-specific specifications and incomplete knowledge of faults and their progression.
Innovation Solution
A hybrid prognostics method combining physics-based and machine learning approaches to identify and encapsulate fundamental degradation modes, allowing for the development of a transferable degradation model that leverages both data-driven and physics-based modeling strengths, integrating sensor data processing and machine learning to estimate remaining useful life.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data-driven approaches are used to develop degradation models, then model accuracy can be improved, but the approach becomes highly dependent on availability of historical data which hampers application in domains where data are not abundantly available
Solution Approach 1:
The patent develops a universal degradation model framework that can be applied across multiple domains (batteries, turbofan engines, lithium-ion batteries) by identifying common fundamental degradation modes (alpha, beta, chi modes) that are shared across different system types. This allows the same model structure to serve multiple purposes without requiring domain-specific redevelopment.
Solution Approach 2:
The patent employs parameter transformation techniques to map domain-specific degradation parameters to a standardized set of fundamental degradation modes. By transforming and normalizing parameters across different domains, the model can adapt to various systems while maintaining consistency in the degradation representation.
2Adaptability or versatility
If physics-based approaches are used to develop degradation models, then the model can be developed without abundant historical data, but the model may struggle with system complexity, nonlinearity, and multi-dimensionality
Solution Approach 1:
The patent segments the complex degradation process into distinct fundamental degradation modes (alpha mode for capacity fade, beta mode for resistance increase, chi mode for structural degradation). This segmentation allows the complex multi-dimensional degradation to be represented as a combination of simpler, physically-interpretable modes, making the model more manageable and applicable to complex systems.
Solution Approach 2:
The patent introduces an intermediary transformation layer that maps complex domain-specific degradation mechanisms to the standardized fundamental degradation modes. This intermediary representation bridges the gap between complex physics-based mechanisms and simplified model representations, enabling accurate modeling without requiring direct representation of all physical complexities.
3Reliability
If domain-specific degradation models are developed for each engineering system, then the model can capture domain-specific specifications, but the model loses transferability and adaptability to different domains
Solution Approach 1:
The patent creates a universal degradation model framework that captures common fundamental degradation modes shared across multiple domains. The model structure is designed to be domain-agnostic, allowing the same fundamental modes (alpha, beta, chi) to represent degradation in batteries, turbofan engines, and other systems, thereby achieving both domain-specific accuracy through parameter customization and cross-domain transferability through shared structure.
Data Source
AI summary
A transferable hybrid method for prognostics of engineering systems based on fundamental degradation modes is provided. The method includes developing a degradation model that represents degradation modes shared in different domains of application through the integration of physics and machine learning. The system measures sensor signals and data processing provides for extracting health indicators correlated with the fundamental degradation modes from sensors data. For the integration of physics and machine learning, the degradation mode is separated into different phases. Before the accelerated degradation phase of a system, the method is looking out to detect when the accelerated phase begins. When accelerated phase is active, the system applies a machine-learning model to provide information on the accelerated degradation phase, and evolves the degradation towards a failure threshold in a simulation of the updated physics-based model to predict the degradation progression. The system estimates the remaining useful life of the target system.