Neural RUL Estimation Using Health Stage Change Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating the Remaining Useful Life (RUL) of systems are inadequate due to the complexity of degradation processes, leading to inaccurate predictions and increased costs from unexpected downtime and maintenance delays.
Innovation Solution
A method utilizing stacked Long Short-Term Memory (LSTM) neural networks and Multi-layer Perceptron (MLP) to analyze historic time series data, determining a health index change point and estimating RUL by training neural networks to minimize differences between multiple RUL estimates, while classifying health stages to suppress false alarms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple neural networks are used to estimate RUL from different perspectives, then estimation accuracy is improved, but system complexity increases
Solution Approach 1:
The RUL estimation system is segmented into multiple specialized neural networks: a first LSTM network for extracting temporal features from sensor data, a second LSTM network for modeling degradation trajectories, and an MLP network for health stage classification. Each network focuses on a specific aspect of degradation analysis, improving overall estimation accuracy while maintaining manageable complexity through functional specialization.
Solution Approach 2:
The patent combines the outputs of multiple neural networks through a unified framework where the LSTM networks generate feature representations that are processed by the MLP classifier. The health stage probabilities from the MLP are integrated with the temporal features from LSTMs to produce a comprehensive RUL estimate, merging different analytical perspectives into a cohesive estimation system.
2Reliability
If health stage classification is implemented to suppress false alarms, then reliability of alerts is improved, but processing time increases
Solution Approach 1:
The system performs preliminary health stage classification using the MLP network before generating RUL alerts. By pre-classifying the current health stage based on sensor data patterns and storing these classifications, the system can quickly retrieve and compare against predicted RUL values without performing full re-analysis, thus suppressing false alarms while minimizing additional processing time.
Solution Approach 2:
The health stage classification acts as an intermediary layer between raw sensor data and RUL estimation. The MLP network transforms complex sensor patterns into simplified health stage probabilities, which then guide the RUL estimation process. This intermediary classification reduces the complexity of direct RUL prediction and filters out false alarm conditions before they propagate to final alerts.
Data Source
AI summary
Systems and methods for determining a remaining useful life of a system. The system and method include one or more processors; a memory coupled to the one or more processors; a data acquisition unit configured to receive run-to-failure time series data; a neural network training unit configured to train a neural network model to determine a point in time that a health index changes from a healthy stage to a degradation stage; a remaining useful life estimation unit configured to estimate a first remaining useful life of the system based on the point in time; estimate a second remaining useful life of the system by converting a feature representation output by the second neural network; minimize the difference between the first remaining useful life and the second remaining useful life; classify the health stage based on a probability; and an output unit configured to send a warning to a user.